video.ipynb

{
 "cells": [
  {
   "cell_type": "markdown",
   "id": "6b069048-25d2-45a8-b43e-a0a9f5f62b92",
   "metadata": {},
   "source": [
    "# OpenAI video testing\n",
    "\n",
    "Conda env: gptwork"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "e47a4ca3-f4dd-435a-844b-212d282dc10d",
   "metadata": {},
   "source": [
    "## Setup"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 2,
   "id": "224e92b0-5cb5-45b9-a062-de4bc307675e",
   "metadata": {},
   "outputs": [],
   "source": [
    "from IPython.display import display, Image, Audio\n",
    "\n",
    "import cv2  # We're using OpenCV to read video, to install !pip install opencv-python\n",
    "import base64\n",
    "import time\n",
    "from openai import OpenAI\n",
    "import os\n",
    "import requests"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 3,
   "id": "bb5fe390-b96d-4ed5-8c50-467d7efed2b9",
   "metadata": {},
   "outputs": [],
   "source": [
    "client = OpenAI(api_key=os.environ.get(\"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\",\n",
    "                                       \"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\"))"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "d2378061-0419-4f0e-b1c7-f58700b2a748",
   "metadata": {},
   "source": [
    "## Functions"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 57,
   "id": "22c74e11-ba42-4c99-b7af-70604fce7ee0",
   "metadata": {},
   "outputs": [],
   "source": [
    "def load_video(video_path, verbose = True):\n",
    "    video = cv2.VideoCapture(video_path)\n",
    "\n",
    "    base64Frames = []\n",
    "    while video.isOpened():\n",
    "        success, frame = video.read()\n",
    "        if not success:\n",
    "            break\n",
    "        _, buffer = cv2.imencode(\".jpg\", frame)\n",
    "        base64Frames.append(base64.b64encode(buffer).decode(\"utf-8\"))\n",
    "    \n",
    "    video.release()\n",
    "    if (verbose):\n",
    "        print(len(base64Frames), \"frames read.\")\n",
    "\n",
    "    return base64Frames\n",
    "\n",
    "def query_video(frames, prompt,\n",
    "                frame_sample = 100,\n",
    "                max_frames = None,\n",
    "                resize_pixels = 768,\n",
    "                max_tokens = 200,\n",
    "                model = \"gpt-4o\",\n",
    "                temperature = 0,\n",
    "                verbose = True):\n",
    "\n",
    "    total_frames = len(frames)\n",
    "    \n",
    "    if max_frames is None:\n",
    "        frame_slice = slice(0, total_frames, frame_sample)\n",
    "    else:\n",
    "        step_size = int(np.ceil(total_frames / max_frames))\n",
    "        frame_slice = slice(0, total_frames, step_size)\n",
    "    \n",
    "    messages = [\n",
    "        { # Prompt\n",
    "            \"role\": \"system\",\n",
    "            \"content\": prompt\n",
    "        },\n",
    "        { # Context in the form of image frames sampled from the video.\n",
    "            \"role\": \"user\",\n",
    "            \"content\": [\n",
    "                *map(lambda x: {\"image\": x, \"resize\": resize_pixels}, frames[frame_slice]),\n",
    "            ],\n",
    "        },\n",
    "    ]\n",
    "    \n",
    "    params = {\n",
    "        \"model\": model,\n",
    "        \"temperature\": temperature,\n",
    "        \"messages\": messages,\n",
    "        \"max_tokens\": max_tokens,\n",
    "    }\n",
    "    \n",
    "    result = client.chat.completions.create(**params)\n",
    "    if verbose:\n",
    "        print(result.choices[0].message.content)\n",
    "    return result"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 48,
   "id": "90398870-3235-4ea8-af70-33ebd3742f77",
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "slice(0, 60, 15)"
      ]
     },
     "execution_count": 48,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "import numpy as np\n",
    "total_frames = 60\n",
    "max_frames = 4\n",
    "\n",
    "frame_slice"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "11cfa703-05b6-40ba-8258-8802ebc5c968",
   "metadata": {},
   "source": [
    "## Prompt tests"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 51,
   "id": "8b1be153-3402-401c-ad51-9a598a22016b",
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "238 frames read.\n"
     ]
    }
   ],
   "source": [
    "# Load a video\n",
    "frames = load_video(\"videos/vaping_modeling.mp4\")"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 63,
   "id": "91e481b3-714b-46e3-bb10-0495fe3977a5",
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Frame 1: Holding necklace - 0% probability of vaping\n",
      "Frame 2: Exhaling smoke - 90% probability of vaping\n",
      "Frame 3: Looking away - 0% probability of vaping\n",
      "Frame 4: Walking forward - 0% probability of vaping\n",
      "Frame 5: Walking forward - 0% probability of vaping\n"
     ]
    }
   ],
   "source": [
    "# Run an Open API GPT query.\n",
    "query_video(frames,\n",
    "            (\"These are frames from a video that I want to upload. Describe what is in this video. \"+\n",
    "            \"Please also provide the numeric probability that there is vaping in each frame. Limit the description to three words per frame\"),\n",
    "           max_frames = 5);"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 60,
   "id": "7c1a21c2-a8e8-4ba2-baa7-f3d159200bba",
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "1. Holding a chain, outdoors. Probability of vaping: [5%]\n",
      "2. Looking away, outdoors. Probability of vaping: [0%]\n",
      "3. Walking, urban setting. Probability of vaping: [0%]\n"
     ]
    }
   ],
   "source": [
    "# Run an Open API GPT query.\n",
    "query_video(frames,\n",
    "            (\"You are an expert in tobacco, e-cigarette, and vaping marketing. These are frames from a video for you to carefully evaluate. \" +\n",
    "             \"Describe what is in this video using no more than 5 words per frame. \"\n",
    "             \"In particular, note any products that are e-cigarettes, tobacco, mods, pods, e-juices, vaping, e-juice containers, or similar. \" +\n",
    "             \"Also provide an accurate numeric probability in brackets that there is vaping in each frame. For example: Probability of vaping: [5%] \"));"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "09b4f9b7-1b42-4bf8-9aa7-591997516bf0",
   "metadata": {},
   "source": [
    "## Run on directory"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 1,
   "id": "100f1916-bcc4-43e9-b16b-328193ce113f",
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Found 102 videos to analyze\n"
     ]
    }
   ],
   "source": [
    "from pathlib import Path\n",
    "video_dir = Path(\"videos/GPT4_themes\")\n",
    "videos = list(video_dir.glob('**/*.mp4'))\n",
    "print(\"Found\", len(videos), \"videos to analyze\")"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 62,
   "id": "de8d76ba-a504-415e-961f-abb207040c25",
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "[PosixPath('videos/GPT4_themes/fashion/fabio_fashion_3.mp4'),\n",
       " PosixPath('videos/GPT4_themes/fashion/chamillioneyes_fashion_1.mp4'),\n",
       " PosixPath('videos/GPT4_themes/fashion/fabio_fashion_2.mp4'),\n",
       " PosixPath('videos/GPT4_themes/fashion/fabio_fashion_18.mp4'),\n",
       " PosixPath('videos/GPT4_themes/fashion/fabio_fashion_19.mp4'),\n",
       " PosixPath('videos/GPT4_themes/fashion/fabio_fashion_14.mp4'),\n",
       " PosixPath('videos/GPT4_themes/fashion/fabio_fashion_4.mp4'),\n",
       " PosixPath('videos/GPT4_themes/fashion/fabio_fashion_17.mp4'),\n",
       " PosixPath('videos/GPT4_themes/fashion/fabio_fashion_8.mp4'),\n",
       " PosixPath('videos/GPT4_themes/fashion/fabio_fashion_9.mp4')]"
      ]
     },
     "execution_count": 62,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "videos[:10]"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 24,
   "id": "fa773c03-9c1c-4b17-ba25-6b141e022b0b",
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "'These are frames from a video that I want to upload. Describe what is in this video. Also describe if there is nicotine vaping in this video. Also describe if there is cannabis vaping in this video. Also describe if there is fashion  in this video. Also describe if there is entertainment (for example, dj, videogames)  in this video. Also describe if there is technology in this video (do not include vaping). Also describe if there is active lifestyle (for example sport) in this video. Limit the description to yes or no responses for each category. Also identify if the person in this video is between the age of 0-11 or 12-17 or 18-21or 22-25 or 26-30 or 31-40 or 41-50 or 51-60 or older than 60 years old. Format the results by surrounding each answer with dollar signs, like $Yes$ or $No$.'"
      ]
     },
     "execution_count": 24,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "prompt_noprobs = (\"These are frames from a video that I want to upload. Describe what is in this video. \"+\n",
    "            \"Also describe if there is nicotine vaping in this video. \"+\n",
    "            \"Also describe if there is cannabis vaping in this video. \"+\n",
    "            \"Also describe if there is fashion (only include modeling) in this video. \"+\n",
    "            \"Also describe if there is entertainment (for example, dj, videogames)  in this video. \"+\n",
    "            \"Also describe if there is active lifestyle (for example sport) in this video. \"+\n",
    "            \"Limit the description to yes or no responses for each category. \" + \n",
    "            \"Also identify if the person in this video is between the age of 0-11 or 12-17 or 18-21\"+ \n",
    "            \"or 22-25 or 26-30 or 31-40 or 41-50 or 51-60 or older than 60 years old. \" +\n",
    "            \"Format the results by surrounding each answer with dollar signs, like $Yes$ or $No$.\")\n",
    "prompt_noprobs"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 322,
   "id": "1485ddc8-86ba-4a02-85df-7f20be5a9c06",
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "'These are frames from a video that I want to upload. Describe what is in this video. Also provide the probability that there is nicotine vaping in this video. Also provide the probability that there is cannabis vaping in this video. Also provide the probability that there is fashion (only include modeling) in this video. Also provide the probability that there is entertainment (for example, dj, videogames)  in this video. Also provide the probability that there is active lifestyle (for example sport) in this video. Limit the description to numeric probability responses for each category, like 100% or 10%. Also provide the probability that the person in this video is between the age of 0-11 or 12-17 or 18-21or 22-25 or 26-30 or 31-40 or 41-50 or 51-60 or older than 60 years old. Format the results by surrounding each numeric probability with dollar signs, like $75%$ or $25%$.'"
      ]
     },
     "execution_count": 322,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "prompt_probs = (\"These are frames from a video that I want to upload. Describe what is in this video. \"+\n",
    "            \"Also provide the probability that there is nicotine vaping in this video. \"+\n",
    "            \"Also provide the probability that there is cannabis vaping in this video. \"+\n",
    "            \"Also provide the probability that there is fashion (only include modeling) in this video. \"+\n",
    "            \"Also provide the probability that there is entertainment (for example, dj, videogames)  in this video. \"+\n",
    "            \"Also provide the probability that there is active lifestyle (for example sport) in this video. \"+\n",
    "            \"Limit the description to numeric probability responses for each category, like 100% or 10%. \" + \n",
    "            \"Also provide the probability that the person in this video is between the age of 0-11 or 12-17 or 18-21\"+ \n",
    "            \"or 22-25 or 26-30 or 31-40 or 41-50 or 51-60 or older than 60 years old. \" +\n",
    "            # Changing 2024-05-20\n",
    "            \"Format the results by surrounding each numeric probability with dollar signs, like $75%$ or $25%$.\")\n",
    "            # Previous version:\n",
    "            #\"Format the results by surrounding each answer with dollar signs, like $75%$ or $25%$.\")\n",
    "prompt_probs"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 64,
   "id": "520ab79c-7288-4189-ad7e-82d41544f32a",
   "metadata": {},
   "outputs": [],
   "source": [
    "def analyze_videos(videos, prompt, max_frames = 4, verbose = False):\n",
    "    results = {}\n",
    "    for video in videos:\n",
    "        video_name = video.stem\n",
    "        if (verbose):\n",
    "            print(f\"\\nAnalyzing {video_name}\")\n",
    "        # Import video\n",
    "        frames = load_video(str(video))\n",
    "        # Run query\n",
    "        result = query_video(frames, prompt, verbose = verbose, max_frames = max_frames)\n",
    "        # Save results\n",
    "        results[video_name] = result\n",
    "    return(results)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 65,
   "id": "6516db6f-92e3-479d-8670-4769fb081625",
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "\n",
      "Analyzing fabio_fashion_3\n",
      "406 frames read.\n",
      "Nicotine vaping: $No$\n",
      "\n",
      "Cannabis vaping: $No$\n",
      "\n",
      "Fashion: $Yes$\n",
      "\n",
      "Entertainment: $No$\n",
      "\n",
      "Technology: $No$\n",
      "\n",
      "Active lifestyle: $No$\n",
      "\n",
      "Age: $26-30$\n",
      "\n",
      "Analyzing chamillioneyes_fashion_1\n",
      "386 frames read.\n",
      "Nicotine vaping: $No$\n",
      "\n",
      "Cannabis vaping: $No$\n",
      "\n",
      "Fashion: $Yes$\n",
      "\n",
      "Entertainment: $No$\n",
      "\n",
      "Technology: $No$\n",
      "\n",
      "Active lifestyle: $No$\n",
      "\n",
      "Age of person: $22-25$\n",
      "\n",
      "Analyzing fabio_fashion_2\n"
     ]
    },
    {
     "ename": "KeyboardInterrupt",
     "evalue": "",
     "output_type": "error",
     "traceback": [
      "\u001b[0;31m---------------------------------------------------------------------------\u001b[0m",
      "\u001b[0;31mKeyboardInterrupt\u001b[0m                         Traceback (most recent call last)",
      "File \u001b[0;32m<timed exec>:1\u001b[0m\n",
      "Cell \u001b[0;32mIn[64], line 8\u001b[0m, in \u001b[0;36manalyze_videos\u001b[0;34m(videos, prompt, max_frames, verbose)\u001b[0m\n\u001b[1;32m      6\u001b[0m     \u001b[38;5;28mprint\u001b[39m(\u001b[38;5;124mf\u001b[39m\u001b[38;5;124m\"\u001b[39m\u001b[38;5;130;01m\\n\u001b[39;00m\u001b[38;5;124mAnalyzing \u001b[39m\u001b[38;5;132;01m{\u001b[39;00mvideo_name\u001b[38;5;132;01m}\u001b[39;00m\u001b[38;5;124m\"\u001b[39m)\n\u001b[1;32m      7\u001b[0m \u001b[38;5;66;03m# Import video\u001b[39;00m\n\u001b[0;32m----> 8\u001b[0m frames \u001b[38;5;241m=\u001b[39m \u001b[43mload_video\u001b[49m\u001b[43m(\u001b[49m\u001b[38;5;28;43mstr\u001b[39;49m\u001b[43m(\u001b[49m\u001b[43mvideo\u001b[49m\u001b[43m)\u001b[49m\u001b[43m)\u001b[49m\n\u001b[1;32m      9\u001b[0m \u001b[38;5;66;03m# Run query\u001b[39;00m\n\u001b[1;32m     10\u001b[0m result \u001b[38;5;241m=\u001b[39m query_video(frames, prompt, verbose \u001b[38;5;241m=\u001b[39m verbose, max_frames \u001b[38;5;241m=\u001b[39m max_frames)\n",
      "Cell \u001b[0;32mIn[57], line 9\u001b[0m, in \u001b[0;36mload_video\u001b[0;34m(video_path, verbose)\u001b[0m\n\u001b[1;32m      7\u001b[0m     \u001b[38;5;28;01mif\u001b[39;00m \u001b[38;5;129;01mnot\u001b[39;00m success:\n\u001b[1;32m      8\u001b[0m         \u001b[38;5;28;01mbreak\u001b[39;00m\n\u001b[0;32m----> 9\u001b[0m     _, buffer \u001b[38;5;241m=\u001b[39m \u001b[43mcv2\u001b[49m\u001b[38;5;241;43m.\u001b[39;49m\u001b[43mimencode\u001b[49m\u001b[43m(\u001b[49m\u001b[38;5;124;43m\"\u001b[39;49m\u001b[38;5;124;43m.jpg\u001b[39;49m\u001b[38;5;124;43m\"\u001b[39;49m\u001b[43m,\u001b[49m\u001b[43m \u001b[49m\u001b[43mframe\u001b[49m\u001b[43m)\u001b[49m\n\u001b[1;32m     10\u001b[0m     base64Frames\u001b[38;5;241m.\u001b[39mappend(base64\u001b[38;5;241m.\u001b[39mb64encode(buffer)\u001b[38;5;241m.\u001b[39mdecode(\u001b[38;5;124m\"\u001b[39m\u001b[38;5;124mutf-8\u001b[39m\u001b[38;5;124m\"\u001b[39m))\n\u001b[1;32m     12\u001b[0m video\u001b[38;5;241m.\u001b[39mrelease()\n",
      "\u001b[0;31mKeyboardInterrupt\u001b[0m: "
     ]
    }
   ],
   "source": [
    "%%time\n",
    "\n",
    "results_noprob = analyze_videos(videos, prompt_noprobs, verbose = True)"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "8fceb4eb-a00b-4eee-b1f4-e5603dacc14f",
   "metadata": {},
   "source": [
    "### Interpret results"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 247,
   "id": "56872ede-9c79-4b68-b05e-71c79d144f9e",
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "[PosixPath('videos/GPT4_themes/fashion/fabio_fashion_3.mp4'),\n",
       " PosixPath('videos/GPT4_themes/fashion/chamillioneyes_fashion_1.mp4'),\n",
       " PosixPath('videos/GPT4_themes/fashion/fabio_fashion_2.mp4'),\n",
       " PosixPath('videos/GPT4_themes/fashion/fabio_fashion_18.mp4'),\n",
       " PosixPath('videos/GPT4_themes/fashion/fabio_fashion_19.mp4')]"
      ]
     },
     "execution_count": 247,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "# Create dataframe with each video and the label theme.\n",
    "videos[:5]"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 2,
   "id": "9207520e-63f7-4f60-a6a4-9268002ea0af",
   "metadata": {},
   "outputs": [],
   "source": [
    "import pandas as pd\n",
    "df = pd.DataFrame(data = {'path': [str(video_i) for video_i in videos]})"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 3,
   "id": "a5e130f3-0c16-4dc9-9620-77d14167405a",
   "metadata": {},
   "outputs": [],
   "source": [
    "df['theme'] = df.path.str.extract(r'(?<=GPT4_themes/)([^/]+)')"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 4,
   "id": "77475ebf-3d62-4d21-8542-a26c63713cce",
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/html": [
       "<div>\n",
       "<style scoped>\n",
       "    .dataframe tbody tr th:only-of-type {\n",
       "        vertical-align: middle;\n",
       "    }\n",
       "\n",
       "    .dataframe tbody tr th {\n",
       "        vertical-align: top;\n",
       "    }\n",
       "\n",
       "    .dataframe thead th {\n",
       "        text-align: right;\n",
       "    }\n",
       "</style>\n",
       "<table border=\"1\" class=\"dataframe\">\n",
       "  <thead>\n",
       "    <tr style=\"text-align: right;\">\n",
       "      <th></th>\n",
       "      <th>path</th>\n",
       "      <th>theme</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>0</th>\n",
       "      <td>videos/GPT4_themes/fashion/fabio_fashion_3.mp4</td>\n",
       "      <td>fashion</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>1</th>\n",
       "      <td>videos/GPT4_themes/fashion/chamillioneyes_fash...</td>\n",
       "      <td>fashion</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>2</th>\n",
       "      <td>videos/GPT4_themes/fashion/fabio_fashion_2.mp4</td>\n",
       "      <td>fashion</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>3</th>\n",
       "      <td>videos/GPT4_themes/fashion/fabio_fashion_18.mp4</td>\n",
       "      <td>fashion</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>4</th>\n",
       "      <td>videos/GPT4_themes/fashion/fabio_fashion_19.mp4</td>\n",
       "      <td>fashion</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>...</th>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>97</th>\n",
       "      <td>videos/GPT4_themes/marijuana/calitrickzz_mj_14...</td>\n",
       "      <td>marijuana</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>98</th>\n",
       "      <td>videos/GPT4_themes/marijuana/calitrickzz_mj_19...</td>\n",
       "      <td>marijuana</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>99</th>\n",
       "      <td>videos/GPT4_themes/marijuana/calitrickzz_mj_4.mp4</td>\n",
       "      <td>marijuana</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>100</th>\n",
       "      <td>videos/GPT4_themes/marijuana/calitrickzz_mj_7.mp4</td>\n",
       "      <td>marijuana</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>101</th>\n",
       "      <td>videos/GPT4_themes/marijuana/calitrickzz_mj_15...</td>\n",
       "      <td>marijuana</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "<p>102 rows × 2 columns</p>\n",
       "</div>"
      ],
      "text/plain": [
       "                                                  path      theme\n",
       "0       videos/GPT4_themes/fashion/fabio_fashion_3.mp4    fashion\n",
       "1    videos/GPT4_themes/fashion/chamillioneyes_fash...    fashion\n",
       "2       videos/GPT4_themes/fashion/fabio_fashion_2.mp4    fashion\n",
       "3      videos/GPT4_themes/fashion/fabio_fashion_18.mp4    fashion\n",
       "4      videos/GPT4_themes/fashion/fabio_fashion_19.mp4    fashion\n",
       "..                                                 ...        ...\n",
       "97   videos/GPT4_themes/marijuana/calitrickzz_mj_14...  marijuana\n",
       "98   videos/GPT4_themes/marijuana/calitrickzz_mj_19...  marijuana\n",
       "99   videos/GPT4_themes/marijuana/calitrickzz_mj_4.mp4  marijuana\n",
       "100  videos/GPT4_themes/marijuana/calitrickzz_mj_7.mp4  marijuana\n",
       "101  videos/GPT4_themes/marijuana/calitrickzz_mj_15...  marijuana\n",
       "\n",
       "[102 rows x 2 columns]"
      ]
     },
     "execution_count": 4,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "df"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 7,
   "id": "730268bc-098f-4148-85ce-d0009b1003e7",
   "metadata": {},
   "outputs": [],
   "source": [
    "df.to_excel(\"themes.xlsx\")"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "id": "8f4d0a04-df02-4073-97e1-dfac6ec3c222",
   "metadata": {},
   "outputs": [],
   "source": [
    "openpyxl"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 5,
   "id": "66e9bd69-ffb9-4c80-a356-b0a1053a4fcb",
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "theme\n",
       "ecigs         22\n",
       "fashion       20\n",
       "health        20\n",
       "technology    20\n",
       "marijuana     20\n",
       "Name: count, dtype: int64"
      ]
     },
     "execution_count": 5,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "df.theme.value_counts()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 327,
   "id": "2f1a155c-b7b7-401c-b55c-2440db341d8b",
   "metadata": {},
   "outputs": [],
   "source": [
    "from itertools import chain\n",
    "\n",
    "def extract_probs(lines):\n",
    "    # This returns a two-dimensional list\n",
    "    string_vals = [re.findall(r'\\$(\\d+)%\\$', line) for line in lines]\n",
    "    onedim_list = list(chain(*string_vals))\n",
    "    list_ints = [int(i) for i in onedim_list]\n",
    "    np_array = np.array(list_ints)\n",
    "    probs = np_array / 100\n",
    "    return(probs)\n",
    "\n",
    "def find_prob_by_keyword(msg, keyword):\n",
    "    prob = None\n",
    "    #  We lowercase the line and the keyword.\n",
    "    target_lines = [line for line in msg.split('\\n') if keyword.lower() in line.lower()]\n",
    "    if len(target_lines) > 0:\n",
    "        probs = extract_probs(target_lines)\n",
    "        if len(probs) > 0:\n",
    "            prob = np.amax(probs)\n",
    "    return(prob)\n",
    "\n",
    "def parse_response(msg):\n",
    "    vaping_prob = find_prob_by_keyword(msg, 'Nicotine vaping')\n",
    "    cannabis_prob = find_prob_by_keyword(msg, 'Cannabis vaping')\n",
    "    fashion_prob = find_prob_by_keyword(msg, 'Fashion')\n",
    "    entertainment_prob = find_prob_by_keyword(msg, 'Entertainment')\n",
    "    lifestyle_prob = find_prob_by_keyword(msg, 'Active lifestyle')\n",
    "    age_0_11 = find_prob_by_keyword(msg, '0-11')\n",
    "    age_12_17 = find_prob_by_keyword(msg, '12-17')\n",
    "    age_18_21 = find_prob_by_keyword(msg, '18-21')\n",
    "    age_22_25 = find_prob_by_keyword(msg, '22-25')\n",
    "    age_26_30 = find_prob_by_keyword(msg, '26-30')\n",
    "    age_31_40 = find_prob_by_keyword(msg, '31-40')\n",
    "    age_41_50 = find_prob_by_keyword(msg, '41-50')\n",
    "    age_51_60 = find_prob_by_keyword(msg, '51-60')\n",
    "    age_61_plus = find_prob_by_keyword(msg, 'Older than 60')\n",
    "    \n",
    "    compiled_probs = {\n",
    "        'nicotine vaping': vaping_prob,\n",
    "        'cannabis vaping': cannabis_prob,\n",
    "        'fashion': fashion_prob,\n",
    "        'entertainment': entertainment_prob,\n",
    "        'lifestyle': lifestyle_prob,\n",
    "        'age_0-11': age_0_11,\n",
    "        'age_12-17': age_12_17,\n",
    "        'age_18-21': age_18_21,\n",
    "        'age_22-25': age_22_25,\n",
    "        'age_26-30': age_26_30,\n",
    "        'age_31-40': age_31_40,\n",
    "        'age_41-50': age_41_50,\n",
    "        'age_51-60': age_51_60,\n",
    "        'age_61_plus': age_61_plus\n",
    "    }\n",
    "\n",
    "    df = pd.DataFrame(compiled_probs, index = [0])\n",
    "    t_df = df.transpose()\n",
    "    t_df.rename(columns = {0: 'prob'}, inplace = True)\n",
    "\n",
    "    missing_probs = sum(t_df.isnull().sum(1))\n",
    "    extracted_probs = t_df.shape[0] - missing_probs\n",
    "\n",
    "    results = {\n",
    "        'df': t_df,\n",
    "        'extracted_probs': extracted_probs\n",
    "    }\n",
    "    return(results)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 325,
   "id": "6e5b18a3-4281-4c3c-beb0-ca1998bef3ad",
   "metadata": {},
   "outputs": [
    {
     "ename": "AttributeError",
     "evalue": "'dict' object has no attribute 'choices'",
     "output_type": "error",
     "traceback": [
      "\u001b[0;31m---------------------------------------------------------------------------\u001b[0m",
      "\u001b[0;31mAttributeError\u001b[0m                            Traceback (most recent call last)",
      "Cell \u001b[0;32mIn[325], line 1\u001b[0m\n\u001b[0;32m----> 1\u001b[0m msg \u001b[38;5;241m=\u001b[39m \u001b[43mresults_probs\u001b[49m\u001b[43m[\u001b[49m\u001b[38;5;124;43m'\u001b[39;49m\u001b[38;5;124;43mfabio_fashion_3\u001b[39;49m\u001b[38;5;124;43m'\u001b[39;49m\u001b[43m]\u001b[49m\u001b[38;5;241;43m.\u001b[39;49m\u001b[43mchoices\u001b[49m[\u001b[38;5;241m0\u001b[39m]\u001b[38;5;241m.\u001b[39mmessage\u001b[38;5;241m.\u001b[39mcontent\n\u001b[1;32m      2\u001b[0m parse_response(msg)\n",
      "\u001b[0;31mAttributeError\u001b[0m: 'dict' object has no attribute 'choices'"
     ]
    }
   ],
   "source": [
    "msg = results_probs['fabio_fashion_3'].choices[0].message.content\n",
    "parse_response(msg)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 208,
   "id": "a34ab5d5-8cf5-453a-86fe-dcb21e70c7d7",
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Analyzing fabio_fashion_3 \n",
      "\n",
      "Nicotine vaping: $0%$\n",
      "Cannabis vaping: $0%$\n",
      "Fashion: $100%$\n",
      "Entertainment: $10%$\n",
      "Technology: $0%$\n",
      "Active lifestyle: $10%$\n",
      "\n",
      "Age group 22-25: $50%$\n",
      "Age group 26-30: $50%$\n",
      "                 prob\n",
      "nicotine vaping   0.0\n",
      "cannabis vaping   0.0\n",
      "fashion           1.0\n",
      "entertainment     0.1\n",
      "lifestyle         0.1\n",
      "age_0-11         None\n",
      "age_12-17        None\n",
      "age_18-21        None\n",
      "age_22-25         0.5\n",
      "age_26-30         0.5\n",
      "age_31-40        None\n",
      "age_41-50        None\n",
      "age_51-60        None\n",
      "age_61_plus      None \n",
      "\n",
      "Analyzing chamillioneyes_fashion_1 \n",
      "\n",
      "Nicotine vaping: $0%$\n",
      "Cannabis vaping: $0%$\n",
      "Fashion: $90%$\n",
      "Entertainment: $20%$\n",
      "Technology: $0%$\n",
      "Active lifestyle: $0%$\n",
      "\n",
      "Age 0-11: $0%$\n",
      "Age 12-17: $0%$\n",
      "Age 18-21: $20%$\n",
      "Age 22-25: $50%$\n",
      "Age 26-30: $30%$\n",
      "Age 31-40: $0%$\n",
      "Age 41-50: $0%$\n",
      "Age 51-60: $0%$\n",
      "Older than 60: $0%$\n",
      "                 prob\n",
      "nicotine vaping   0.0\n",
      "cannabis vaping   0.0\n",
      "fashion           0.9\n",
      "entertainment     0.2\n",
      "lifestyle         0.0\n",
      "age_0-11          0.0\n",
      "age_12-17         0.0\n",
      "age_18-21         0.2\n",
      "age_22-25         0.5\n",
      "age_26-30         0.3\n",
      "age_31-40         0.0\n",
      "age_41-50         0.0\n",
      "age_51-60         0.0\n",
      "age_61_plus      None \n",
      "\n",
      "Analyzing fabio_fashion_2 \n",
      "\n",
      "$Nicotine vaping: 0%$\n",
      "\n",
      "$Cannabis vaping: 0%$\n",
      "\n",
      "$Fashion: 100%$\n",
      "\n",
      "$Entertainment: 0%$\n",
      "\n",
      "$Technology: 0%$\n",
      "\n",
      "$Active lifestyle: 10%$\n",
      "\n",
      "$Age 0-11: 0%$\n",
      "\n",
      "$Age 12-17: 0%$\n",
      "\n",
      "$Age 18-21: 10%$\n",
      "\n",
      "$Age 22-25: 30%$\n",
      "\n",
      "$Age 26-30: 40%$\n",
      "\n",
      "$Age 31-40: 20%$\n",
      "\n",
      "$Age 41-50: 0%$\n",
      "\n",
      "$Age 51-60: 0%$\n",
      "\n",
      "$Age older than 60: 0%$\n",
      "Retrying... only found 0 probabilities.\n",
      "\n",
      "Retrying... only found 0 probabilities.\n",
      "\n",
      "Retrying... only found 0 probabilities.\n",
      "\n",
      "                 prob\n",
      "nicotine vaping  None\n",
      "cannabis vaping  None\n",
      "fashion          None\n",
      "entertainment    None\n",
      "lifestyle        None\n",
      "age_0-11         None\n",
      "age_12-17        None\n",
      "age_18-21        None\n",
      "age_22-25        None\n",
      "age_26-30        None\n",
      "age_31-40        None\n",
      "age_41-50        None\n",
      "age_51-60        None\n",
      "age_61_plus      None \n",
      "\n",
      "Analyzing fabio_fashion_18 \n",
      "\n",
      "Nicotine vaping: $0%$\n",
      "Cannabis vaping: $0%$\n",
      "Fashion: $100%$\n",
      "Entertainment: $10%$\n",
      "Technology: $0%$\n",
      "Active lifestyle: $10%$\n",
      "\n",
      "Age 0-11: $0%$\n",
      "Age 12-17: $0%$\n",
      "Age 18-21: $10%$\n",
      "Age 22-25: $50%$\n",
      "Age 26-30: $30%$\n",
      "Age 31-40: $10%$\n",
      "Age 41-50: $0%$\n",
      "Age 51-60: $0%$\n",
      "Older than 60: $0%$\n",
      "                 prob\n",
      "nicotine vaping   0.0\n",
      "cannabis vaping   0.0\n",
      "fashion           1.0\n",
      "entertainment     0.1\n",
      "lifestyle         0.1\n",
      "age_0-11          0.0\n",
      "age_12-17         0.0\n",
      "age_18-21         0.1\n",
      "age_22-25         0.5\n",
      "age_26-30         0.3\n",
      "age_31-40         0.1\n",
      "age_41-50         0.0\n",
      "age_51-60         0.0\n",
      "age_61_plus      None \n",
      "\n",
      "Analyzing fabio_fashion_19 \n",
      "\n",
      "Nicotine vaping: $0%$\n",
      "Cannabis vaping: $0%$\n",
      "Fashion: $100%$\n",
      "Entertainment: $10%$\n",
      "Technology: $0%$\n",
      "Active lifestyle: $10%$\n",
      "\n",
      "Age group (12-17): $0%$\n",
      "Age group (18-21): $10%$\n",
      "Age group (22-25): $50%$\n",
      "Age group (26-30): $30%$\n",
      "Age group (31-40): $10%$\n",
      "                 prob\n",
      "nicotine vaping   0.0\n",
      "cannabis vaping   0.0\n",
      "fashion           1.0\n",
      "entertainment     0.1\n",
      "lifestyle         0.1\n",
      "age_0-11         None\n",
      "age_12-17         0.0\n",
      "age_18-21         0.1\n",
      "age_22-25         0.5\n",
      "age_26-30         0.3\n",
      "age_31-40         0.1\n",
      "age_41-50        None\n",
      "age_51-60        None\n",
      "age_61_plus      None \n",
      "\n"
     ]
    }
   ],
   "source": [
    "max_tries = 3\n",
    "for video, data in results_probs.items():\n",
    "    print(\"Analyzing\", video, \"\\n\")\n",
    "    msg = data.choices[0].message.content\n",
    "    print(msg)\n",
    "    tries = 0\n",
    "    while tries < max_tries:\n",
    "        tries = tries + 1\n",
    "        result = parse_response(msg)\n",
    "        if result['extracted_probs'] > 5:\n",
    "            break\n",
    "        print(\"Retrying... only found\", result['extracted_probs'], \"probabilities.\\n\")\n",
    "        \n",
    "    print(result['df'], \"\\n\")"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "4382e6f5-5a55-410f-9584-97696f481a68",
   "metadata": {},
   "source": [
    "## Run and parse simultaneously"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 333,
   "id": "3009088e-7f6c-45bc-bb18-0bd174382978",
   "metadata": {},
   "outputs": [],
   "source": [
    "from openai import RateLimitError\n",
    "\n",
    "def analyze_videos_probs(videos, prompt, max_frames = 4, min_tries = 1, max_tries = 3, min_extracted_probs = 5, verbose = False):\n",
    "    results = {}\n",
    "\n",
    "    stop_execution = False\n",
    "    for video in videos:\n",
    "        video_name = video.stem\n",
    "        if (verbose):\n",
    "            print(f\"\\nAnalyzing {video_name}\")\n",
    "        # Import video\n",
    "        frames = load_video(str(video))\n",
    "\n",
    "        tries = 0\n",
    "\n",
    "        while tries < max_tries:\n",
    "            tries = tries + 1\n",
    "\n",
    "            parsed_result = None\n",
    "            num_probs = 0\n",
    "            query_result = None\n",
    "            msg = None\n",
    "            \n",
    "            # Run query\n",
    "            try:\n",
    "                query_result = query_video(frames, prompt, verbose = verbose, max_frames = max_frames)\n",
    "                msg = query_result.choices[0].message.content\n",
    "                parsed_result = parse_response(msg)\n",
    "                if parsed_result is not None:\n",
    "                    num_probs = parsed_result['extracted_probs']\n",
    "            except RateLimitError as e:\n",
    "                print(\"Ran into rate limit exception:\")\n",
    "                print(e)\n",
    "                print(\"Stopping execution\")\n",
    "                stop_execution = True\n",
    "                break\n",
    "            except BadRequestError as e:\n",
    "                print(\"Ran into Bad Request error:\")\n",
    "                print(e)\n",
    "            except Exception as e:\n",
    "                print(\"Catching a novel exception:\")\n",
    "                print(e)\n",
    "\n",
    "            if num_probs >= min_extracted_probs:\n",
    "                break\n",
    "            \n",
    "            print(\"Retrying... only found\", num_probs, \"probabilities.\\n\")\n",
    "\n",
    "        if verbose and not stop_execution:\n",
    "            print(parsed_result)\n",
    "            \n",
    "        # Save results\n",
    "        results[video_name] = {\n",
    "            'query': query_result,\n",
    "            'msg': msg,\n",
    "            'parsed': parsed_result,\n",
    "            'tries': tries\n",
    "        }\n",
    "\n",
    "        if stop_execution:\n",
    "            break\n",
    "    return(results)        \n",
    "    "
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 334,
   "id": "fd814b41-fa0e-4c28-b4f2-8f3f972ad499",
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "\n",
      "Analyzing fabio_fashion_3\n",
      "406 frames read.\n",
      "Nicotine vaping: $5%$\n",
      "Cannabis vaping: $5%$\n",
      "Fashion (modeling): $80%$\n",
      "Entertainment: $10%$\n",
      "Active lifestyle (sport): $10%$\n",
      "\n",
      "Age group probabilities:\n",
      "0-11 years old: $0%$\n",
      "12-17 years old: $5%$\n",
      "18-21 years old: $20%$\n",
      "22-25 years old: $40%$\n",
      "26-30 years old: $25%$\n",
      "31-40 years old: $10%$\n",
      "41-50 years old: $0%$\n",
      "51-60 years old: $0%$\n",
      "Older than 60 years old: $0%$\n",
      "{'df':                  prob\n",
      "nicotine vaping  0.05\n",
      "cannabis vaping  0.05\n",
      "fashion          0.80\n",
      "entertainment    0.10\n",
      "lifestyle        0.10\n",
      "age_0-11         0.00\n",
      "age_12-17        0.05\n",
      "age_18-21        0.20\n",
      "age_22-25        0.40\n",
      "age_26-30        0.25\n",
      "age_31-40        0.10\n",
      "age_41-50        0.00\n",
      "age_51-60        0.00\n",
      "age_61_plus      0.00, 'extracted_probs': 14}\n",
      "\n",
      "Analyzing chamillioneyes_fashion_1\n",
      "386 frames read.\n",
      "Nicotine vaping: $0%$\n",
      "Cannabis vaping: $0%$\n",
      "Fashion (modeling): $80%$\n",
      "Entertainment: $20%$\n",
      "Active lifestyle: $10%$\n",
      "\n",
      "Age 0-11: $0%$\n",
      "Age 12-17: $0%$\n",
      "Age 18-21: $20%$\n",
      "Age 22-25: $50%$\n",
      "Age 26-30: $20%$\n",
      "Age 31-40: $10%$\n",
      "Age 41-50: $0%$\n",
      "Age 51-60: $0%$\n",
      "Older than 60: $0%$\n",
      "{'df':                  prob\n",
      "nicotine vaping   0.0\n",
      "cannabis vaping   0.0\n",
      "fashion           0.8\n",
      "entertainment     0.2\n",
      "lifestyle         0.1\n",
      "age_0-11          0.0\n",
      "age_12-17         0.0\n",
      "age_18-21         0.2\n",
      "age_22-25         0.5\n",
      "age_26-30         0.2\n",
      "age_31-40         0.1\n",
      "age_41-50         0.0\n",
      "age_51-60         0.0\n",
      "age_61_plus       0.0, 'extracted_probs': 14}\n",
      "\n",
      "Analyzing fabio_fashion_2\n",
      "436 frames read.\n",
      "Nicotine vaping: $0%$\n",
      "Cannabis vaping: $0%$\n",
      "Fashion (modeling): $80%$\n",
      "Entertainment: $10%$\n",
      "Active lifestyle (sport): $10%$\n",
      "\n",
      "Age group probabilities:\n",
      "- 0-11 years old: $0%$\n",
      "- 12-17 years old: $0%$\n",
      "- 18-21 years old: $10%$\n",
      "- 22-25 years old: $30%$\n",
      "- 26-30 years old: $40%$\n",
      "- 31-40 years old: $20%$\n",
      "- 41-50 years old: $0%$\n",
      "- 51-60 years old: $0%$\n",
      "- Older than 60 years old: $0%$\n",
      "{'df':                  prob\n",
      "nicotine vaping   0.0\n",
      "cannabis vaping   0.0\n",
      "fashion           0.8\n",
      "entertainment     0.1\n",
      "lifestyle         0.1\n",
      "age_0-11          0.0\n",
      "age_12-17         0.0\n",
      "age_18-21         0.1\n",
      "age_22-25         0.3\n",
      "age_26-30         0.4\n",
      "age_31-40         0.2\n",
      "age_41-50         0.0\n",
      "age_51-60         0.0\n",
      "age_61_plus       0.0, 'extracted_probs': 14}\n",
      "\n",
      "Analyzing fabio_fashion_18\n",
      "237 frames read.\n",
      "Nicotine vaping: $0%$\n",
      "Cannabis vaping: $0%$\n",
      "Fashion (modeling): $80%$\n",
      "Entertainment (dj, videogames): $10%$\n",
      "Active lifestyle (sport): $10%$\n",
      "\n",
      "Age 0-11: $0%$\n",
      "Age 12-17: $0%$\n",
      "Age 18-21: $10%$\n",
      "Age 22-25: $40%$\n",
      "Age 26-30: $30%$\n",
      "Age 31-40: $20%$\n",
      "Age 41-50: $0%$\n",
      "Age 51-60: $0%$\n",
      "Older than 60: $0%$\n",
      "{'df':                  prob\n",
      "nicotine vaping   0.0\n",
      "cannabis vaping   0.0\n",
      "fashion           0.8\n",
      "entertainment     0.1\n",
      "lifestyle         0.1\n",
      "age_0-11          0.0\n",
      "age_12-17         0.0\n",
      "age_18-21         0.1\n",
      "age_22-25         0.4\n",
      "age_26-30         0.3\n",
      "age_31-40         0.2\n",
      "age_41-50         0.0\n",
      "age_51-60         0.0\n",
      "age_61_plus       0.0, 'extracted_probs': 14}\n",
      "\n",
      "Analyzing fabio_fashion_19\n",
      "251 frames read.\n",
      "Nicotine vaping: $0%$\n",
      "Cannabis vaping: $0%$\n",
      "Fashion (modeling): $80%$\n",
      "Entertainment: $10%$\n",
      "Active lifestyle: $10%$\n",
      "\n",
      "Age 0-11: $0%$\n",
      "Age 12-17: $0%$\n",
      "Age 18-21: $10%$\n",
      "Age 22-25: $40%$\n",
      "Age 26-30: $40%$\n",
      "Age 31-40: $10%$\n",
      "Age 41-50: $0%$\n",
      "Age 51-60: $0%$\n",
      "Older than 60: $0%$\n",
      "{'df':                  prob\n",
      "nicotine vaping   0.0\n",
      "cannabis vaping   0.0\n",
      "fashion           0.8\n",
      "entertainment     0.1\n",
      "lifestyle         0.1\n",
      "age_0-11          0.0\n",
      "age_12-17         0.0\n",
      "age_18-21         0.1\n",
      "age_22-25         0.4\n",
      "age_26-30         0.4\n",
      "age_31-40         0.1\n",
      "age_41-50         0.0\n",
      "age_51-60         0.0\n",
      "age_61_plus       0.0, 'extracted_probs': 14}\n",
      "\n",
      "Analyzing fabio_fashion_14\n",
      "219 frames read.\n",
      "Nicotine vaping: $0%$\n",
      "Cannabis vaping: $0%$\n",
      "Fashion (modeling): $90%$\n",
      "Entertainment: $10%$\n",
      "Active lifestyle: $10%$\n",
      "\n",
      "Age 0-11: $0%$\n",
      "Age 12-17: $0%$\n",
      "Age 18-21: $20%$\n",
      "Age 22-25: $50%$\n",
      "Age 26-30: $20%$\n",
      "Age 31-40: $10%$\n",
      "Age 41-50: $0%$\n",
      "Age 51-60: $0%$\n",
      "Older than 60: $0%$\n",
      "{'df':                  prob\n",
      "nicotine vaping   0.0\n",
      "cannabis vaping   0.0\n",
      "fashion           0.9\n",
      "entertainment     0.1\n",
      "lifestyle         0.1\n",
      "age_0-11          0.0\n",
      "age_12-17         0.0\n",
      "age_18-21         0.2\n",
      "age_22-25         0.5\n",
      "age_26-30         0.2\n",
      "age_31-40         0.1\n",
      "age_41-50         0.0\n",
      "age_51-60         0.0\n",
      "age_61_plus       0.0, 'extracted_probs': 14}\n",
      "\n",
      "Analyzing fabio_fashion_4\n",
      "503 frames read.\n",
      "Nicotine vaping: $10%$\n",
      "Cannabis vaping: $10%$\n",
      "Fashion (modeling): $80%$\n",
      "Entertainment: $20%$\n",
      "Active lifestyle: $20%$\n",
      "\n",
      "Age 0-11: $0%$\n",
      "Age 12-17: $0%$\n",
      "Age 18-21: $20%$\n",
      "Age 22-25: $50%$\n",
      "Age 26-30: $20%$\n",
      "Age 31-40: $10%$\n",
      "Age 41-50: $0%$\n",
      "Age 51-60: $0%$\n",
      "Older than 60: $0%$\n",
      "{'df':                  prob\n",
      "nicotine vaping   0.1\n",
      "cannabis vaping   0.1\n",
      "fashion           0.8\n",
      "entertainment     0.2\n",
      "lifestyle         0.2\n",
      "age_0-11          0.0\n",
      "age_12-17         0.0\n",
      "age_18-21         0.2\n",
      "age_22-25         0.5\n",
      "age_26-30         0.2\n",
      "age_31-40         0.1\n",
      "age_41-50         0.0\n",
      "age_51-60         0.0\n",
      "age_61_plus       0.0, 'extracted_probs': 14}\n",
      "\n",
      "Analyzing fabio_fashion_17\n",
      "261 frames read.\n",
      "Nicotine vaping: $0%$\n",
      "Cannabis vaping: $0%$\n",
      "Fashion (modeling): $80%$\n",
      "Entertainment: $10%$\n",
      "Active lifestyle (sport): $10%$\n",
      "\n",
      "Age group probabilities:\n",
      "0-11 years old: $0%$\n",
      "12-17 years old: $0%$\n",
      "18-21 years old: $20%$\n",
      "22-25 years old: $50%$\n",
      "26-30 years old: $20%$\n",
      "31-40 years old: $10%$\n",
      "41-50 years old: $0%$\n",
      "51-60 years old: $0%$\n",
      "Older than 60 years old: $0%$\n",
      "{'df':                  prob\n",
      "nicotine vaping   0.0\n",
      "cannabis vaping   0.0\n",
      "fashion           0.8\n",
      "entertainment     0.1\n",
      "lifestyle         0.1\n",
      "age_0-11          0.0\n",
      "age_12-17         0.0\n",
      "age_18-21         0.2\n",
      "age_22-25         0.5\n",
      "age_26-30         0.2\n",
      "age_31-40         0.1\n",
      "age_41-50         0.0\n",
      "age_51-60         0.0\n",
      "age_61_plus       0.0, 'extracted_probs': 14}\n",
      "\n",
      "Analyzing fabio_fashion_8\n",
      "443 frames read.\n",
      "- Nicotine vaping: $0%$\n",
      "- Cannabis vaping: $0%$\n",
      "- Fashion (modeling): $80%$\n",
      "- Entertainment: $10%$\n",
      "- Active lifestyle (sport): $10%$\n",
      "\n",
      "- Age 0-11: $0%$\n",
      "- Age 12-17: $0%$\n",
      "- Age 18-21: $20%$\n",
      "- Age 22-25: $50%$\n",
      "- Age 26-30: $20%$\n",
      "- Age 31-40: $10%$\n",
      "- Age 41-50: $0%$\n",
      "- Age 51-60: $0%$\n",
      "- Older than 60: $0%$\n",
      "{'df':                  prob\n",
      "nicotine vaping   0.0\n",
      "cannabis vaping   0.0\n",
      "fashion           0.8\n",
      "entertainment     0.1\n",
      "lifestyle         0.1\n",
      "age_0-11          0.0\n",
      "age_12-17         0.0\n",
      "age_18-21         0.2\n",
      "age_22-25         0.5\n",
      "age_26-30         0.2\n",
      "age_31-40         0.1\n",
      "age_41-50         0.0\n",
      "age_51-60         0.0\n",
      "age_61_plus       0.0, 'extracted_probs': 14}\n",
      "\n",
      "Analyzing fabio_fashion_9\n",
      "400 frames read.\n",
      "Nicotine vaping: $0%$\n",
      "Cannabis vaping: $0%$\n",
      "Fashion (modeling): $90%$\n",
      "Entertainment: $10%$\n",
      "Active lifestyle: $10%$\n",
      "\n",
      "Age group probabilities:\n",
      "- 0-11 years old: $0%$\n",
      "- 12-17 years old: $0%$\n",
      "- 18-21 years old: $20%$\n",
      "- 22-25 years old: $50%$\n",
      "- 26-30 years old: $20%$\n",
      "- 31-40 years old: $10%$\n",
      "- 41-50 years old: $0%$\n",
      "- 51-60 years old: $0%$\n",
      "- Older than 60 years old: $0%$\n",
      "{'df':                  prob\n",
      "nicotine vaping   0.0\n",
      "cannabis vaping   0.0\n",
      "fashion           0.9\n",
      "entertainment     0.1\n",
      "lifestyle         0.1\n",
      "age_0-11          0.0\n",
      "age_12-17         0.0\n",
      "age_18-21         0.2\n",
      "age_22-25         0.5\n",
      "age_26-30         0.2\n",
      "age_31-40         0.1\n",
      "age_41-50         0.0\n",
      "age_51-60         0.0\n",
      "age_61_plus       0.0, 'extracted_probs': 14}\n",
      "\n",
      "Analyzing fabio_fashion_12\n",
      "434 frames read.\n",
      "Nicotine vaping: $0%$\n",
      "Cannabis vaping: $0%$\n",
      "Fashion (modeling): $80%$\n",
      "Entertainment: $10%$\n",
      "Active lifestyle: $70%$\n",
      "\n",
      "Age group probabilities:\n",
      "0-11 years old: $0%$\n",
      "12-17 years old: $0%$\n",
      "18-21 years old: $10%$\n",
      "22-25 years old: $30%$\n",
      "26-30 years old: $40%$\n",
      "31-40 years old: $20%$\n",
      "41-50 years old: $0%$\n",
      "51-60 years old: $0%$\n",
      "Older than 60 years old: $0%$\n",
      "{'df':                  prob\n",
      "nicotine vaping   0.0\n",
      "cannabis vaping   0.0\n",
      "fashion           0.8\n",
      "entertainment     0.1\n",
      "lifestyle         0.7\n",
      "age_0-11          0.0\n",
      "age_12-17         0.0\n",
      "age_18-21         0.1\n",
      "age_22-25         0.3\n",
      "age_26-30         0.4\n",
      "age_31-40         0.2\n",
      "age_41-50         0.0\n",
      "age_51-60         0.0\n",
      "age_61_plus       0.0, 'extracted_probs': 14}\n",
      "\n",
      "Analyzing fabio_fashion_1\n",
      "224 frames read.\n",
      "Nicotine vaping: $0%$\n",
      "Cannabis vaping: $0%$\n",
      "Fashion (modeling): $80%$\n",
      "Entertainment: $10%$\n",
      "Active lifestyle: $10%$\n",
      "\n",
      "Age probabilities:\n",
      "0-11 years old: $0%$\n",
      "12-17 years old: $0%$\n",
      "18-21 years old: $10%$\n",
      "22-25 years old: $30%$\n",
      "26-30 years old: $40%$\n",
      "31-40 years old: $20%$\n",
      "41-50 years old: $0%$\n",
      "51-60 years old: $0%$\n",
      "Older than 60 years old: $0%$\n",
      "{'df':                  prob\n",
      "nicotine vaping   0.0\n",
      "cannabis vaping   0.0\n",
      "fashion           0.8\n",
      "entertainment     0.1\n",
      "lifestyle         0.1\n",
      "age_0-11          0.0\n",
      "age_12-17         0.0\n",
      "age_18-21         0.1\n",
      "age_22-25         0.3\n",
      "age_26-30         0.4\n",
      "age_31-40         0.2\n",
      "age_41-50         0.0\n",
      "age_51-60         0.0\n",
      "age_61_plus       0.0, 'extracted_probs': 14}\n",
      "\n",
      "Analyzing fabio_fashion_16\n",
      "195 frames read.\n",
      "Nicotine vaping: $5%$\n",
      "Cannabis vaping: $5%$\n",
      "Fashion (modeling): $90%$\n",
      "Entertainment: $10%$\n",
      "Active lifestyle: $10%$\n",
      "\n",
      "Age 0-11: $0%$\n",
      "Age 12-17: $0%$\n",
      "Age 18-21: $10%$\n",
      "Age 22-25: $40%$\n",
      "Age 26-30: $40%$\n",
      "Age 31-40: $10%$\n",
      "Age 41-50: $0%$\n",
      "Age 51-60: $0%$\n",
      "Older than 60: $0%$\n",
      "{'df':                  prob\n",
      "nicotine vaping  0.05\n",
      "cannabis vaping  0.05\n",
      "fashion          0.90\n",
      "entertainment    0.10\n",
      "lifestyle        0.10\n",
      "age_0-11         0.00\n",
      "age_12-17        0.00\n",
      "age_18-21        0.10\n",
      "age_22-25        0.40\n",
      "age_26-30        0.40\n",
      "age_31-40        0.10\n",
      "age_41-50        0.00\n",
      "age_51-60        0.00\n",
      "age_61_plus      0.00, 'extracted_probs': 14}\n",
      "\n",
      "Analyzing fabio_fashion_11\n",
      "498 frames read.\n",
      "Nicotine vaping: $0%$\n",
      "Cannabis vaping: $0%$\n",
      "Fashion (modeling): $90%$\n",
      "Entertainment: $10%$\n",
      "Active lifestyle (sport): $20%$\n",
      "\n",
      "Age probabilities:\n",
      "0-11 years old: $0%$\n",
      "12-17 years old: $0%$\n",
      "18-21 years old: $20%$\n",
      "22-25 years old: $40%$\n",
      "26-30 years old: $30%$\n",
      "31-40 years old: $10%$\n",
      "41-50 years old: $0%$\n",
      "51-60 years old: $0%$\n",
      "Older than 60 years old: $0%$\n",
      "{'df':                  prob\n",
      "nicotine vaping   0.0\n",
      "cannabis vaping   0.0\n",
      "fashion           0.9\n",
      "entertainment     0.1\n",
      "lifestyle         0.2\n",
      "age_0-11          0.0\n",
      "age_12-17         0.0\n",
      "age_18-21         0.2\n",
      "age_22-25         0.4\n",
      "age_26-30         0.3\n",
      "age_31-40         0.1\n",
      "age_41-50         0.0\n",
      "age_51-60         0.0\n",
      "age_61_plus       0.0, 'extracted_probs': 14}\n",
      "\n",
      "Analyzing fabio_fashion_7\n",
      "320 frames read.\n",
      "Nicotine vaping: $0%$\n",
      "Cannabis vaping: $0%$\n",
      "Fashion (modeling): $80%$\n",
      "Entertainment: $10%$\n",
      "Active lifestyle: $10%$\n",
      "\n",
      "Age 0-11: $0%$\n",
      "Age 12-17: $0%$\n",
      "Age 18-21: $10%$\n",
      "Age 22-25: $40%$\n",
      "Age 26-30: $40%$\n",
      "Age 31-40: $10%$\n",
      "Age 41-50: $0%$\n",
      "Age 51-60: $0%$\n",
      "Older than 60: $0%$\n",
      "{'df':                  prob\n",
      "nicotine vaping   0.0\n",
      "cannabis vaping   0.0\n",
      "fashion           0.8\n",
      "entertainment     0.1\n",
      "lifestyle         0.1\n",
      "age_0-11          0.0\n",
      "age_12-17         0.0\n",
      "age_18-21         0.1\n",
      "age_22-25         0.4\n",
      "age_26-30         0.4\n",
      "age_31-40         0.1\n",
      "age_41-50         0.0\n",
      "age_51-60         0.0\n",
      "age_61_plus       0.0, 'extracted_probs': 14}\n",
      "\n",
      "Analyzing fabio_fashion_15\n",
      "240 frames read.\n",
      "Nicotine vaping: $0%$\n",
      "Cannabis vaping: $0%$\n",
      "Fashion (modeling): $80%$\n",
      "Entertainment: $10%$\n",
      "Active lifestyle (sport): $10%$\n",
      "\n",
      "Age 0-11: $0%$\n",
      "Age 12-17: $0%$\n",
      "Age 18-21: $10%$\n",
      "Age 22-25: $30%$\n",
      "Age 26-30: $40%$\n",
      "Age 31-40: $20%$\n",
      "Age 41-50: $0%$\n",
      "Age 51-60: $0%$\n",
      "Older than 60: $0%$\n",
      "{'df':                  prob\n",
      "nicotine vaping   0.0\n",
      "cannabis vaping   0.0\n",
      "fashion           0.8\n",
      "entertainment     0.1\n",
      "lifestyle         0.1\n",
      "age_0-11          0.0\n",
      "age_12-17         0.0\n",
      "age_18-21         0.1\n",
      "age_22-25         0.3\n",
      "age_26-30         0.4\n",
      "age_31-40         0.2\n",
      "age_41-50         0.0\n",
      "age_51-60         0.0\n",
      "age_61_plus       0.0, 'extracted_probs': 14}\n",
      "\n",
      "Analyzing fabio_fashion_6\n",
      "183 frames read.\n",
      "Nicotine vaping: $5%$\n",
      "Cannabis vaping: $5%$\n",
      "Fashion (modeling): $80%$\n",
      "Entertainment: $10%$\n",
      "Active lifestyle: $10%$\n",
      "\n",
      "Age:\n",
      "- 0-11 years old: $0%$\n",
      "- 12-17 years old: $0%$\n",
      "- 18-21 years old: $10%$\n",
      "- 22-25 years old: $30%$\n",
      "- 26-30 years old: $40%$\n",
      "- 31-40 years old: $20%$\n",
      "- 41-50 years old: $0%$\n",
      "- 51-60 years old: $0%$\n",
      "- Older than 60 years old: $0%$\n",
      "{'df':                  prob\n",
      "nicotine vaping  0.05\n",
      "cannabis vaping  0.05\n",
      "fashion          0.80\n",
      "entertainment    0.10\n",
      "lifestyle        0.10\n",
      "age_0-11         0.00\n",
      "age_12-17        0.00\n",
      "age_18-21        0.10\n",
      "age_22-25        0.30\n",
      "age_26-30        0.40\n",
      "age_31-40        0.20\n",
      "age_41-50        0.00\n",
      "age_51-60        0.00\n",
      "age_61_plus      0.00, 'extracted_probs': 14}\n",
      "\n",
      "Analyzing fabio_fashion_13\n",
      "208 frames read.\n",
      "Nicotine vaping: $0%$\n",
      "Cannabis vaping: $0%$\n",
      "Fashion (modeling): $90%$\n",
      "Entertainment: $10%$\n",
      "Active lifestyle (sport): $10%$\n",
      "\n",
      "Age range probabilities:\n",
      "- 0-11 years old: $0%$\n",
      "- 12-17 years old: $0%$\n",
      "- 18-21 years old: $10%$\n",
      "- 22-25 years old: $30%$\n",
      "- 26-30 years old: $40%$\n",
      "- 31-40 years old: $20%$\n",
      "- 41-50 years old: $0%$\n",
      "- 51-60 years old: $0%$\n",
      "- Older than 60 years old: $0%$\n",
      "{'df':                  prob\n",
      "nicotine vaping   0.0\n",
      "cannabis vaping   0.0\n",
      "fashion           0.9\n",
      "entertainment     0.1\n",
      "lifestyle         0.1\n",
      "age_0-11          0.0\n",
      "age_12-17         0.0\n",
      "age_18-21         0.1\n",
      "age_22-25         0.3\n",
      "age_26-30         0.4\n",
      "age_31-40         0.2\n",
      "age_41-50         0.0\n",
      "age_51-60         0.0\n",
      "age_61_plus       0.0, 'extracted_probs': 14}\n",
      "\n",
      "Analyzing fabio_fashion_10_andpod\n",
      "248 frames read.\n",
      "Nicotine vaping: $10%$\n",
      "Cannabis vaping: $10%$\n",
      "Fashion (modeling): $60%$\n",
      "Entertainment: $20%$\n",
      "Active lifestyle: $20%$\n",
      "\n",
      "Age 0-11: $0%$\n",
      "Age 12-17: $0%$\n",
      "Age 18-21: $20%$\n",
      "Age 22-25: $40%$\n",
      "Age 26-30: $30%$\n",
      "Age 31-40: $10%$\n",
      "Age 41-50: $0%$\n",
      "Age 51-60: $0%$\n",
      "Older than 60: $0%$\n",
      "{'df':                  prob\n",
      "nicotine vaping   0.1\n",
      "cannabis vaping   0.1\n",
      "fashion           0.6\n",
      "entertainment     0.2\n",
      "lifestyle         0.2\n",
      "age_0-11          0.0\n",
      "age_12-17         0.0\n",
      "age_18-21         0.2\n",
      "age_22-25         0.4\n",
      "age_26-30         0.3\n",
      "age_31-40         0.1\n",
      "age_41-50         0.0\n",
      "age_51-60         0.0\n",
      "age_61_plus       0.0, 'extracted_probs': 14}\n",
      "\n",
      "Analyzing fabio_fashion_5\n",
      "461 frames read.\n",
      "Nicotine vaping: $0%$\n",
      "Cannabis vaping: $0%$\n",
      "Fashion (modeling): $80%$\n",
      "Entertainment: $10%$\n",
      "Active lifestyle: $10%$\n",
      "\n",
      "Age group probabilities:\n",
      "- 0-11 years old: $0%$\n",
      "- 12-17 years old: $0%$\n",
      "- 18-21 years old: $10%$\n",
      "- 22-25 years old: $30%$\n",
      "- 26-30 years old: $40%$\n",
      "- 31-40 years old: $20%$\n",
      "- 41-50 years old: $0%$\n",
      "- 51-60 years old: $0%$\n",
      "- Older than 60 years old: $0%$\n",
      "{'df':                  prob\n",
      "nicotine vaping   0.0\n",
      "cannabis vaping   0.0\n",
      "fashion           0.8\n",
      "entertainment     0.1\n",
      "lifestyle         0.1\n",
      "age_0-11          0.0\n",
      "age_12-17         0.0\n",
      "age_18-21         0.1\n",
      "age_22-25         0.3\n",
      "age_26-30         0.4\n",
      "age_31-40         0.2\n",
      "age_41-50         0.0\n",
      "age_51-60         0.0\n",
      "age_61_plus       0.0, 'extracted_probs': 14}\n",
      "\n",
      "Analyzing fabio_health_3\n",
      "1730 frames read.\n",
      "Nicotine vaping: $0%$\n",
      "Cannabis vaping: $0%$\n",
      "Fashion (modeling): $0%$\n",
      "Entertainment: $0%$\n",
      "Active lifestyle (sport): $100%$\n",
      "\n",
      "Age group probabilities:\n",
      "0-11 years old: $0%$\n",
      "12-17 years old: $0%$\n",
      "18-21 years old: $10%$\n",
      "22-25 years old: $30%$\n",
      "26-30 years old: $40%$\n",
      "31-40 years old: $20%$\n",
      "41-50 years old: $0%$\n",
      "51-60 years old: $0%$\n",
      "Older than 60 years old: $0%$\n",
      "{'df':                  prob\n",
      "nicotine vaping   0.0\n",
      "cannabis vaping   0.0\n",
      "fashion           0.0\n",
      "entertainment     0.0\n",
      "lifestyle         1.0\n",
      "age_0-11          0.0\n",
      "age_12-17         0.0\n",
      "age_18-21         0.1\n",
      "age_22-25         0.3\n",
      "age_26-30         0.4\n",
      "age_31-40         0.2\n",
      "age_41-50         0.0\n",
      "age_51-60         0.0\n",
      "age_61_plus       0.0, 'extracted_probs': 14}\n",
      "\n",
      "Analyzing drewdirps_health_2\n",
      "1800 frames read.\n",
      "Nicotine vaping: $0%$\n",
      "Cannabis vaping: $0%$\n",
      "Fashion (modeling): $0%$\n",
      "Entertainment: $0%$\n",
      "Active lifestyle (sport): $100%$\n",
      "\n",
      "Age group probabilities:\n",
      "0-11 years old: $0%$\n",
      "12-17 years old: $0%$\n",
      "18-21 years old: $10%$\n",
      "22-25 years old: $20%$\n",
      "26-30 years old: $30%$\n",
      "31-40 years old: $30%$\n",
      "41-50 years old: $10%$\n",
      "51-60 years old: $0%$\n",
      "Older than 60 years old: $0%$\n",
      "{'df':                  prob\n",
      "nicotine vaping   0.0\n",
      "cannabis vaping   0.0\n",
      "fashion           0.0\n",
      "entertainment     0.0\n",
      "lifestyle         1.0\n",
      "age_0-11          0.0\n",
      "age_12-17         0.0\n",
      "age_18-21         0.1\n",
      "age_22-25         0.2\n",
      "age_26-30         0.3\n",
      "age_31-40         0.3\n",
      "age_41-50         0.1\n",
      "age_51-60         0.0\n",
      "age_61_plus       0.0, 'extracted_probs': 14}\n",
      "\n",
      "Analyzing fabio_health_2_and_pod\n",
      "1802 frames read.\n",
      "Nicotine vaping: $90%$\n",
      "Cannabis vaping: $10%$\n",
      "Fashion (modeling): $50%$\n",
      "Entertainment: $20%$\n",
      "Active lifestyle: $30%$\n",
      "\n",
      "Age 0-11: $0%$\n",
      "Age 12-17: $10%$\n",
      "Age 18-21: $20%$\n",
      "Age 22-25: $30%$\n",
      "Age 26-30: $20%$\n",
      "Age 31-40: $10%$\n",
      "Age 41-50: $5%$\n",
      "Age 51-60: $5%$\n",
      "Older than 60: $0%$\n",
      "{'df':                  prob\n",
      "nicotine vaping  0.90\n",
      "cannabis vaping  0.10\n",
      "fashion          0.50\n",
      "entertainment    0.20\n",
      "lifestyle        0.30\n",
      "age_0-11         0.00\n",
      "age_12-17        0.10\n",
      "age_18-21        0.20\n",
      "age_22-25        0.30\n",
      "age_26-30        0.20\n",
      "age_31-40        0.10\n",
      "age_41-50        0.05\n",
      "age_51-60        0.05\n",
      "age_61_plus      0.00, 'extracted_probs': 14}\n",
      "\n",
      "Analyzing arabella_health_1\n",
      "1671 frames read.\n",
      "Nicotine vaping: $0%$\n",
      "Cannabis vaping: $0%$\n",
      "Fashion (modeling): $0%$\n",
      "Entertainment: $0%$\n",
      "Active lifestyle: $100%$\n",
      "\n",
      "Age group probabilities:\n",
      "- 0-11 years old: $0%$\n",
      "- 12-17 years old: $0%$\n",
      "- 18-21 years old: $20%$\n",
      "- 22-25 years old: $40%$\n",
      "- 26-30 years old: $30%$\n",
      "- 31-40 years old: $10%$\n",
      "- 41-50 years old: $0%$\n",
      "- 51-60 years old: $0%$\n",
      "- Older than 60 years old: $0%$\n",
      "{'df':                  prob\n",
      "nicotine vaping   0.0\n",
      "cannabis vaping   0.0\n",
      "fashion           0.0\n",
      "entertainment     0.0\n",
      "lifestyle         1.0\n",
      "age_0-11          0.0\n",
      "age_12-17         0.0\n",
      "age_18-21         0.2\n",
      "age_22-25         0.4\n",
      "age_26-30         0.3\n",
      "age_31-40         0.1\n",
      "age_41-50         0.0\n",
      "age_51-60         0.0\n",
      "age_61_plus       0.0, 'extracted_probs': 14}\n",
      "\n",
      "Analyzing chamil_health_vape\n",
      "263 frames read.\n",
      "Nicotine vaping: $90%$\n",
      "Cannabis vaping: $10%$\n",
      "Fashion (modeling): $0%$\n",
      "Entertainment: $0%$\n",
      "Active lifestyle: $100%$\n",
      "\n",
      "Age group probabilities:\n",
      "- 0-11 years old: $0%$\n",
      "- 12-17 years old: $0%$\n",
      "- 18-21 years old: $10%$\n",
      "- 22-25 years old: $30%$\n",
      "- 26-30 years old: $40%$\n",
      "- 31-40 years old: $20%$\n",
      "- 41-50 years old: $0%$\n",
      "- 51-60 years old: $0%$\n",
      "- Older than 60 years old: $0%$\n",
      "{'df':                  prob\n",
      "nicotine vaping   0.9\n",
      "cannabis vaping   0.1\n",
      "fashion           0.0\n",
      "entertainment     0.0\n",
      "lifestyle         1.0\n",
      "age_0-11          0.0\n",
      "age_12-17         0.0\n",
      "age_18-21         0.1\n",
      "age_22-25         0.3\n",
      "age_26-30         0.4\n",
      "age_31-40         0.2\n",
      "age_41-50         0.0\n",
      "age_51-60         0.0\n",
      "age_61_plus       0.0, 'extracted_probs': 14}\n",
      "\n",
      "Analyzing chamillioneyes_health_1\n",
      "2127 frames read.\n",
      "Nicotine vaping: $0%$\n",
      "Cannabis vaping: $0%$\n",
      "Fashion (modeling): $0%$\n",
      "Entertainment: $0%$\n",
      "Active lifestyle: $100%$\n",
      "\n",
      "Age group probabilities:\n",
      "- 0-11 years old: $0%$\n",
      "- 12-17 years old: $0%$\n",
      "- 18-21 years old: $10%$\n",
      "- 22-25 years old: $30%$\n",
      "- 26-30 years old: $40%$\n",
      "- 31-40 years old: $20%$\n",
      "- 41-50 years old: $0%$\n",
      "- 51-60 years old: $0%$\n",
      "- Older than 60 years old: $0%$\n",
      "{'df':                  prob\n",
      "nicotine vaping   0.0\n",
      "cannabis vaping   0.0\n",
      "fashion           0.0\n",
      "entertainment     0.0\n",
      "lifestyle         1.0\n",
      "age_0-11          0.0\n",
      "age_12-17         0.0\n",
      "age_18-21         0.1\n",
      "age_22-25         0.3\n",
      "age_26-30         0.4\n",
      "age_31-40         0.2\n",
      "age_41-50         0.0\n",
      "age_51-60         0.0\n",
      "age_61_plus       0.0, 'extracted_probs': 14}\n",
      "\n",
      "Analyzing edripps_health_vape\n",
      "247 frames read.\n",
      "- Probability of nicotine vaping: $90%$\n",
      "- Probability of cannabis vaping: $10%$\n",
      "- Probability of fashion (modeling): $0%$\n",
      "- Probability of entertainment: $10%$\n",
      "- Probability of active lifestyle (sport): $80%$\n",
      "- Probability of age 0-11: $0%$\n",
      "- Probability of age 12-17: $0%$\n",
      "- Probability of age 18-21: $20%$\n",
      "- Probability of age 22-25: $50%$\n",
      "- Probability of age 26-30: $30%$\n",
      "- Probability of age 31-40: $0%$\n",
      "- Probability of age 41-50: $0%$\n",
      "- Probability of age 51-60: $0%$\n",
      "- Probability of older than 60 years old: $0%$\n",
      "{'df':                  prob\n",
      "nicotine vaping   0.9\n",
      "cannabis vaping   0.1\n",
      "fashion           0.0\n",
      "entertainment     0.1\n",
      "lifestyle         0.8\n",
      "age_0-11          0.0\n",
      "age_12-17         0.0\n",
      "age_18-21         0.2\n",
      "age_22-25         0.5\n",
      "age_26-30         0.3\n",
      "age_31-40         0.0\n",
      "age_41-50         0.0\n",
      "age_51-60         0.0\n",
      "age_61_plus       0.0, 'extracted_probs': 14}\n",
      "\n",
      "Analyzing drewdirps_health_4\n",
      "269 frames read.\n",
      "Nicotine vaping: $0%$\n",
      "Cannabis vaping: $0%$\n",
      "Fashion (modeling): $0%$\n",
      "Entertainment: $0%$\n",
      "Active lifestyle (sport): $100%$\n",
      "\n",
      "Age probability:\n",
      "- 0-11 years old: $0%$\n",
      "- 12-17 years old: $0%$\n",
      "- 18-21 years old: $10%$\n",
      "- 22-25 years old: $30%$\n",
      "- 26-30 years old: $40%$\n",
      "- 31-40 years old: $20%$\n",
      "- 41-50 years old: $0%$\n",
      "- 51-60 years old: $0%$\n",
      "- Older than 60 years old: $0%$\n",
      "{'df':                  prob\n",
      "nicotine vaping   0.0\n",
      "cannabis vaping   0.0\n",
      "fashion           0.0\n",
      "entertainment     0.0\n",
      "lifestyle         1.0\n",
      "age_0-11          0.0\n",
      "age_12-17         0.0\n",
      "age_18-21         0.1\n",
      "age_22-25         0.3\n",
      "age_26-30         0.4\n",
      "age_31-40         0.2\n",
      "age_41-50         0.0\n",
      "age_51-60         0.0\n",
      "age_61_plus       0.0, 'extracted_probs': 14}\n",
      "\n",
      "Analyzing drewdirps_health_1\n",
      "1417 frames read.\n",
      "Nicotine vaping: $0%$\n",
      "Cannabis vaping: $0%$\n",
      "Fashion (modeling): $0%$\n",
      "Entertainment: $0%$\n",
      "Active lifestyle: $100%$\n",
      "\n",
      "Age 0-11: $0%$\n",
      "Age 12-17: $0%$\n",
      "Age 18-21: $10%$\n",
      "Age 22-25: $30%$\n",
      "Age 26-30: $40%$\n",
      "Age 31-40: $20%$\n",
      "Age 41-50: $0%$\n",
      "Age 51-60: $0%$\n",
      "Older than 60: $0%$\n",
      "{'df':                  prob\n",
      "nicotine vaping   0.0\n",
      "cannabis vaping   0.0\n",
      "fashion           0.0\n",
      "entertainment     0.0\n",
      "lifestyle         1.0\n",
      "age_0-11          0.0\n",
      "age_12-17         0.0\n",
      "age_18-21         0.1\n",
      "age_22-25         0.3\n",
      "age_26-30         0.4\n",
      "age_31-40         0.2\n",
      "age_41-50         0.0\n",
      "age_51-60         0.0\n",
      "age_61_plus       0.0, 'extracted_probs': 14}\n",
      "\n",
      "Analyzing vapes_aby_health_1_andvape\n",
      "1290 frames read.\n",
      "- Probability of nicotine vaping: $90%$\n",
      "- Probability of cannabis vaping: $10%$\n",
      "- Probability of fashion (modeling): $10%$\n",
      "- Probability of entertainment: $20%$\n",
      "- Probability of active lifestyle (sport): $100%$\n",
      "\n",
      "- Probability of age 0-11: $0%$\n",
      "- Probability of age 12-17: $20%$\n",
      "- Probability of age 18-21: $50%$\n",
      "- Probability of age 22-25: $20%$\n",
      "- Probability of age 26-30: $10%$\n",
      "- Probability of age 31-40: $0%$\n",
      "- Probability of age 41-50: $0%$\n",
      "- Probability of age 51-60: $0%$\n",
      "- Probability of older than 60 years old: $0%$\n",
      "{'df':                  prob\n",
      "nicotine vaping   0.9\n",
      "cannabis vaping   0.1\n",
      "fashion           0.1\n",
      "entertainment     0.2\n",
      "lifestyle         1.0\n",
      "age_0-11          0.0\n",
      "age_12-17         0.2\n",
      "age_18-21         0.5\n",
      "age_22-25         0.2\n",
      "age_26-30         0.1\n",
      "age_31-40         0.0\n",
      "age_41-50         0.0\n",
      "age_51-60         0.0\n",
      "age_61_plus       0.0, 'extracted_probs': 14}\n",
      "\n",
      "Analyzing vapes_aby_health_2_andvape\n",
      "605 frames read.\n",
      "- Probability of nicotine vaping: $0%$\n",
      "- Probability of cannabis vaping: $0%$\n",
      "- Probability of fashion (modeling): $0%$\n",
      "- Probability of entertainment: $10%$\n",
      "- Probability of active lifestyle (sport): $100%$\n",
      "\n",
      "- Probability of age 0-11: $0%$\n",
      "- Probability of age 12-17: $50%$\n",
      "- Probability of age 18-21: $50%$\n",
      "- Probability of age 22-25: $0%$\n",
      "- Probability of age 26-30: $0%$\n",
      "- Probability of age 31-40: $0%$\n",
      "- Probability of age 41-50: $0%$\n",
      "- Probability of age 51-60: $0%$\n",
      "- Probability of older than 60 years old: $0%$\n",
      "{'df':                  prob\n",
      "nicotine vaping   0.0\n",
      "cannabis vaping   0.0\n",
      "fashion           0.0\n",
      "entertainment     0.1\n",
      "lifestyle         1.0\n",
      "age_0-11          0.0\n",
      "age_12-17         0.5\n",
      "age_18-21         0.5\n",
      "age_22-25         0.0\n",
      "age_26-30         0.0\n",
      "age_31-40         0.0\n",
      "age_41-50         0.0\n",
      "age_51-60         0.0\n",
      "age_61_plus       0.0, 'extracted_probs': 14}\n",
      "\n",
      "Analyzing alex_health_vape\n",
      "238 frames read.\n",
      "- Probability of nicotine vaping: $90%$\n",
      "- Probability of cannabis vaping: $10%$\n",
      "- Probability of fashion (modeling): $20%$\n",
      "- Probability of entertainment: $10%$\n",
      "- Probability of active lifestyle (sport): $80%$\n",
      "\n",
      "- Probability of age 0-11: $0%$\n",
      "- Probability of age 12-17: $0%$\n",
      "- Probability of age 18-21: $10%$\n",
      "- Probability of age 22-25: $30%$\n",
      "- Probability of age 26-30: $40%$\n",
      "- Probability of age 31-40: $20%$\n",
      "- Probability of age 41-50: $0%$\n",
      "- Probability of age 51-60: $0%$\n",
      "- Probability of older than 60 years old: $0%$\n",
      "{'df':                  prob\n",
      "nicotine vaping   0.9\n",
      "cannabis vaping   0.1\n",
      "fashion           0.2\n",
      "entertainment     0.1\n",
      "lifestyle         0.8\n",
      "age_0-11          0.0\n",
      "age_12-17         0.0\n",
      "age_18-21         0.1\n",
      "age_22-25         0.3\n",
      "age_26-30         0.4\n",
      "age_31-40         0.2\n",
      "age_41-50         0.0\n",
      "age_51-60         0.0\n",
      "age_61_plus       0.0, 'extracted_probs': 14}\n",
      "\n",
      "Analyzing drew_health_vape\n",
      "273 frames read.\n",
      "Nicotine vaping: $90%$\n",
      "Cannabis vaping: $50%$\n",
      "Fashion (modeling): $0%$\n",
      "Entertainment: $10%$\n",
      "Active lifestyle: $80%$\n",
      "\n",
      "Age group probabilities:\n",
      "- 0-11 years old: $0%$\n",
      "- 12-17 years old: $0%$\n",
      "- 18-21 years old: $10%$\n",
      "- 22-25 years old: $30%$\n",
      "- 26-30 years old: $40%$\n",
      "- 31-40 years old: $20%$\n",
      "- 41-50 years old: $0%$\n",
      "- 51-60 years old: $0%$\n",
      "- Older than 60 years old: $0%$\n",
      "{'df':                  prob\n",
      "nicotine vaping   0.9\n",
      "cannabis vaping   0.5\n",
      "fashion           0.0\n",
      "entertainment     0.1\n",
      "lifestyle         0.8\n",
      "age_0-11          0.0\n",
      "age_12-17         0.0\n",
      "age_18-21         0.1\n",
      "age_22-25         0.3\n",
      "age_26-30         0.4\n",
      "age_31-40         0.2\n",
      "age_41-50         0.0\n",
      "age_51-60         0.0\n",
      "age_61_plus       0.0, 'extracted_probs': 14}\n",
      "\n",
      "Analyzing chamillioneyes_health_2\n",
      "770 frames read.\n",
      "- Probability of nicotine vaping: $0%$\n",
      "- Probability of cannabis vaping: $0%$\n",
      "- Probability of fashion (modeling): $0%$\n",
      "- Probability of entertainment: $0%$\n",
      "- Probability of active lifestyle (sport): $100%$\n",
      "\n",
      "- Probability of age 0-11: $0%$\n",
      "- Probability of age 12-17: $0%$\n",
      "- Probability of age 18-21: $10%$\n",
      "- Probability of age 22-25: $40%$\n",
      "- Probability of age 26-30: $40%$\n",
      "- Probability of age 31-40: $10%$\n",
      "- Probability of age 41-50: $0%$\n",
      "- Probability of age 51-60: $0%$\n",
      "- Probability of older than 60 years old: $0%$\n",
      "{'df':                  prob\n",
      "nicotine vaping   0.0\n",
      "cannabis vaping   0.0\n",
      "fashion           0.0\n",
      "entertainment     0.0\n",
      "lifestyle         1.0\n",
      "age_0-11          0.0\n",
      "age_12-17         0.0\n",
      "age_18-21         0.1\n",
      "age_22-25         0.4\n",
      "age_26-30         0.4\n",
      "age_31-40         0.1\n",
      "age_41-50         0.0\n",
      "age_51-60         0.0\n",
      "age_61_plus       0.0, 'extracted_probs': 14}\n",
      "\n",
      "Analyzing vapes_aby_health_vape\n",
      "237 frames read.\n",
      "Nicotine vaping: $100%$\n",
      "Cannabis vaping: $10%$\n",
      "Fashion (modeling): $10%$\n",
      "Entertainment: $20%$\n",
      "Active lifestyle: $50%$\n",
      "\n",
      "Age group probabilities:\n",
      "- 0-11 years old: $0%$\n",
      "- 12-17 years old: $10%$\n",
      "- 18-21 years old: $20%$\n",
      "- 22-25 years old: $40%$\n",
      "- 26-30 years old: $20%$\n",
      "- 31-40 years old: $10%$\n",
      "- 41-50 years old: $0%$\n",
      "- 51-60 years old: $0%$\n",
      "- Older than 60 years old: $0%$\n",
      "{'df':                  prob\n",
      "nicotine vaping   1.0\n",
      "cannabis vaping   0.1\n",
      "fashion           0.1\n",
      "entertainment     0.2\n",
      "lifestyle         0.5\n",
      "age_0-11          0.0\n",
      "age_12-17         0.1\n",
      "age_18-21         0.2\n",
      "age_22-25         0.4\n",
      "age_26-30         0.2\n",
      "age_31-40         0.1\n",
      "age_41-50         0.0\n",
      "age_51-60         0.0\n",
      "age_61_plus       0.0, 'extracted_probs': 14}\n",
      "\n",
      "Analyzing fabio_health_1\n",
      "434 frames read.\n",
      "- Probability of nicotine vaping: $0%$\n",
      "- Probability of cannabis vaping: $0%$\n",
      "- Probability of fashion (modeling): $80%$\n",
      "- Probability of entertainment: $10%$\n",
      "- Probability of active lifestyle (sport): $70%$\n",
      "\n",
      "- Probability of age 0-11: $0%$\n",
      "- Probability of age 12-17: $0%$\n",
      "- Probability of age 18-21: $10%$\n",
      "- Probability of age 22-25: $30%$\n",
      "- Probability of age 26-30: $40%$\n",
      "- Probability of age 31-40: $20%$\n",
      "- Probability of age 41-50: $0%$\n",
      "- Probability of age 51-60: $0%$\n",
      "- Probability of older than 60 years old: $0%$\n",
      "{'df':                  prob\n",
      "nicotine vaping   0.0\n",
      "cannabis vaping   0.0\n",
      "fashion           0.8\n",
      "entertainment     0.1\n",
      "lifestyle         0.7\n",
      "age_0-11          0.0\n",
      "age_12-17         0.0\n",
      "age_18-21         0.1\n",
      "age_22-25         0.3\n",
      "age_26-30         0.4\n",
      "age_31-40         0.2\n",
      "age_41-50         0.0\n",
      "age_51-60         0.0\n",
      "age_61_plus       0.0, 'extracted_probs': 14}\n",
      "\n",
      "Analyzing drewdirps_health_3\n",
      "1346 frames read.\n",
      "Nicotine vaping: $0%$\n",
      "Cannabis vaping: $0%$\n",
      "Fashion (modeling): $0%$\n",
      "Entertainment: $0%$\n",
      "Active lifestyle (sport): $100%$\n",
      "\n",
      "Age group probabilities:\n",
      "0-11 years old: $0%$\n",
      "12-17 years old: $0%$\n",
      "18-21 years old: $10%$\n",
      "22-25 years old: $30%$\n",
      "26-30 years old: $40%$\n",
      "31-40 years old: $20%$\n",
      "41-50 years old: $0%$\n",
      "51-60 years old: $0%$\n",
      "Older than 60 years old: $0%$\n",
      "{'df':                  prob\n",
      "nicotine vaping   0.0\n",
      "cannabis vaping   0.0\n",
      "fashion           0.0\n",
      "entertainment     0.0\n",
      "lifestyle         1.0\n",
      "age_0-11          0.0\n",
      "age_12-17         0.0\n",
      "age_18-21         0.1\n",
      "age_22-25         0.3\n",
      "age_26-30         0.4\n",
      "age_31-40         0.2\n",
      "age_41-50         0.0\n",
      "age_51-60         0.0\n",
      "age_61_plus       0.0, 'extracted_probs': 14}\n",
      "\n",
      "Analyzing alex_health_2\n",
      "431 frames read.\n",
      "Nicotine vaping: $0%$\n",
      "Cannabis vaping: $0%$\n",
      "Fashion (modeling): $0%$\n",
      "Entertainment: $0%$\n",
      "Active lifestyle: $100%$\n",
      "\n",
      "Age group probabilities:\n",
      "0-11 years old: $0%$\n",
      "12-17 years old: $0%$\n",
      "18-21 years old: $10%$\n",
      "22-25 years old: $30%$\n",
      "26-30 years old: $40%$\n",
      "31-40 years old: $20%$\n",
      "41-50 years old: $0%$\n",
      "51-60 years old: $0%$\n",
      "Older than 60 years old: $0%$\n",
      "{'df':                  prob\n",
      "nicotine vaping   0.0\n",
      "cannabis vaping   0.0\n",
      "fashion           0.0\n",
      "entertainment     0.0\n",
      "lifestyle         1.0\n",
      "age_0-11          0.0\n",
      "age_12-17         0.0\n",
      "age_18-21         0.1\n",
      "age_22-25         0.3\n",
      "age_26-30         0.4\n",
      "age_31-40         0.2\n",
      "age_41-50         0.0\n",
      "age_51-60         0.0\n",
      "age_61_plus       0.0, 'extracted_probs': 14}\n",
      "\n",
      "Analyzing arabella_health_2\n",
      "2241 frames read.\n",
      "Nicotine vaping: $100%$\n",
      "Cannabis vaping: $0%$\n",
      "Fashion (modeling): $0%$\n",
      "Entertainment: $0%$\n",
      "Active lifestyle (sport): $100%$\n",
      "\n",
      "Age group probabilities:\n",
      "0-11 years old: $0%$\n",
      "12-17 years old: $0%$\n",
      "18-21 years old: $20%$\n",
      "22-25 years old: $50%$\n",
      "26-30 years old: $30%$\n",
      "31-40 years old: $0%$\n",
      "41-50 years old: $0%$\n",
      "51-60 years old: $0%$\n",
      "Older than 60 years old: $0%$\n",
      "{'df':                  prob\n",
      "nicotine vaping   1.0\n",
      "cannabis vaping   0.0\n",
      "fashion           0.0\n",
      "entertainment     0.0\n",
      "lifestyle         1.0\n",
      "age_0-11          0.0\n",
      "age_12-17         0.0\n",
      "age_18-21         0.2\n",
      "age_22-25         0.5\n",
      "age_26-30         0.3\n",
      "age_31-40         0.0\n",
      "age_41-50         0.0\n",
      "age_51-60         0.0\n",
      "age_61_plus       0.0, 'extracted_probs': 14}\n",
      "\n",
      "Analyzing alex_health_1\n",
      "737 frames read.\n",
      "Nicotine vaping: $0%$\n",
      "Cannabis vaping: $0%$\n",
      "Fashion (modeling): $10%$\n",
      "Entertainment: $0%$\n",
      "Active lifestyle: $100%$\n",
      "\n",
      "Age group probabilities:\n",
      "0-11 years old: $0%$\n",
      "12-17 years old: $0%$\n",
      "18-21 years old: $10%$\n",
      "22-25 years old: $40%$\n",
      "26-30 years old: $30%$\n",
      "31-40 years old: $20%$\n",
      "41-50 years old: $0%$\n",
      "51-60 years old: $0%$\n",
      "Older than 60 years old: $0%$\n",
      "{'df':                  prob\n",
      "nicotine vaping   0.0\n",
      "cannabis vaping   0.0\n",
      "fashion           0.1\n",
      "entertainment     0.0\n",
      "lifestyle         1.0\n",
      "age_0-11          0.0\n",
      "age_12-17         0.0\n",
      "age_18-21         0.1\n",
      "age_22-25         0.4\n",
      "age_26-30         0.3\n",
      "age_31-40         0.2\n",
      "age_41-50         0.0\n",
      "age_51-60         0.0\n",
      "age_61_plus       0.0, 'extracted_probs': 14}\n",
      "\n",
      "Analyzing drewdirps_ecigs_3\n",
      "722 frames read.\n",
      "- Probability of nicotine vaping: $90%$\n",
      "- Probability of cannabis vaping: $30%$\n",
      "- Probability of fashion (modeling): $0%$\n",
      "- Probability of entertainment (e.g., DJ, video games): $10%$\n",
      "- Probability of active lifestyle (e.g., sport): $0%$\n",
      "\n",
      "- Probability of age 0-11: $0%$\n",
      "- Probability of age 12-17: $0%$\n",
      "- Probability of age 18-21: $10%$\n",
      "- Probability of age 22-25: $30%$\n",
      "- Probability of age 26-30: $40%$\n",
      "- Probability of age 31-40: $20%$\n",
      "- Probability of age 41-50: $0%$\n",
      "- Probability of age 51-60: $0%$\n",
      "- Probability of older than 60 years old: $0%$\n",
      "{'df':                  prob\n",
      "nicotine vaping   0.9\n",
      "cannabis vaping   0.3\n",
      "fashion           0.0\n",
      "entertainment     0.1\n",
      "lifestyle         0.0\n",
      "age_0-11          0.0\n",
      "age_12-17         0.0\n",
      "age_18-21         0.1\n",
      "age_22-25         0.3\n",
      "age_26-30         0.4\n",
      "age_31-40         0.2\n",
      "age_41-50         0.0\n",
      "age_51-60         0.0\n",
      "age_61_plus       0.0, 'extracted_probs': 14}\n",
      "\n",
      "Analyzing chamillioneyes_ecgis_1\n",
      "573 frames read.\n",
      "Nicotine vaping: $90%$\n",
      "Cannabis vaping: $10%$\n",
      "Fashion (modeling): $0%$\n",
      "Entertainment: $0%$\n",
      "Active lifestyle (sport): $0%$\n",
      "\n",
      "Age range 22-25: $60%$\n",
      "Age range 26-30: $40%$\n",
      "{'df':                  prob\n",
      "nicotine vaping   0.9\n",
      "cannabis vaping   0.1\n",
      "fashion           0.0\n",
      "entertainment     0.0\n",
      "lifestyle         0.0\n",
      "age_0-11         None\n",
      "age_12-17        None\n",
      "age_18-21        None\n",
      "age_22-25         0.6\n",
      "age_26-30         0.4\n",
      "age_31-40        None\n",
      "age_41-50        None\n",
      "age_51-60        None\n",
      "age_61_plus      None, 'extracted_probs': 7}\n",
      "\n",
      "Analyzing drewdirps_ecigs_4\n",
      "374 frames read.\n",
      "Nicotine vaping: $100%$\n",
      "Cannabis vaping: $0%$\n",
      "Fashion (modeling): $0%$\n",
      "Entertainment: $0%$\n",
      "Active lifestyle (sport): $0%$\n",
      "\n",
      "Age group probabilities:\n",
      "0-11 years old: $0%$\n",
      "12-17 years old: $0%$\n",
      "18-21 years old: $10%$\n",
      "22-25 years old: $30%$\n",
      "26-30 years old: $40%$\n",
      "31-40 years old: $20%$\n",
      "41-50 years old: $0%$\n",
      "51-60 years old: $0%$\n",
      "Older than 60 years old: $0%$\n",
      "{'df':                  prob\n",
      "nicotine vaping   1.0\n",
      "cannabis vaping   0.0\n",
      "fashion           0.0\n",
      "entertainment     0.0\n",
      "lifestyle         0.0\n",
      "age_0-11          0.0\n",
      "age_12-17         0.0\n",
      "age_18-21         0.1\n",
      "age_22-25         0.3\n",
      "age_26-30         0.4\n",
      "age_31-40         0.2\n",
      "age_41-50         0.0\n",
      "age_51-60         0.0\n",
      "age_61_plus       0.0, 'extracted_probs': 14}\n",
      "\n",
      "Analyzing edripss_ecgis_2\n",
      "557 frames read.\n",
      "Nicotine vaping: $100%$\n",
      "Cannabis vaping: $10%$\n",
      "Fashion (modeling): $0%$\n",
      "Entertainment: $20%$\n",
      "Active lifestyle (sport): $0%$\n",
      "\n",
      "Age 0-11: $0%$\n",
      "Age 12-17: $0%$\n",
      "Age 18-21: $20%$\n",
      "Age 22-25: $50%$\n",
      "Age 26-30: $20%$\n",
      "Age 31-40: $10%$\n",
      "Age 41-50: $0%$\n",
      "Age 51-60: $0%$\n",
      "Older than 60: $0%$\n",
      "{'df':                  prob\n",
      "nicotine vaping   1.0\n",
      "cannabis vaping   0.1\n",
      "fashion           0.0\n",
      "entertainment     0.2\n",
      "lifestyle         0.0\n",
      "age_0-11          0.0\n",
      "age_12-17         0.0\n",
      "age_18-21         0.2\n",
      "age_22-25         0.5\n",
      "age_26-30         0.2\n",
      "age_31-40         0.1\n",
      "age_41-50         0.0\n",
      "age_51-60         0.0\n",
      "age_61_plus       0.0, 'extracted_probs': 14}\n",
      "\n",
      "Analyzing chamillioneyes_ecigs_2\n",
      "320 frames read.\n",
      "Nicotine vaping: $0%$\n",
      "Cannabis vaping: $0%$\n",
      "Fashion (modeling): $0%$\n",
      "Entertainment: $0%$\n",
      "Active lifestyle (sport): $0%$\n",
      "\n",
      "Age probability:\n",
      "- 0-11 years old: $0%$\n",
      "- 12-17 years old: $0%$\n",
      "- 18-21 years old: $10%$\n",
      "- 22-25 years old: $30%$\n",
      "- 26-30 years old: $40%$\n",
      "- 31-40 years old: $20%$\n",
      "- 41-50 years old: $0%$\n",
      "- 51-60 years old: $0%$\n",
      "- Older than 60 years old: $0%$\n",
      "{'df':                  prob\n",
      "nicotine vaping   0.0\n",
      "cannabis vaping   0.0\n",
      "fashion           0.0\n",
      "entertainment     0.0\n",
      "lifestyle         0.0\n",
      "age_0-11          0.0\n",
      "age_12-17         0.0\n",
      "age_18-21         0.1\n",
      "age_22-25         0.3\n",
      "age_26-30         0.4\n",
      "age_31-40         0.2\n",
      "age_41-50         0.0\n",
      "age_51-60         0.0\n",
      "age_61_plus       0.0, 'extracted_probs': 14}\n",
      "\n",
      "Analyzing calitrickzz_ecigs_3\n",
      "943 frames read.\n",
      "Nicotine vaping: $100%$\n",
      "Cannabis vaping: $0%$\n",
      "Fashion (modeling): $0%$\n",
      "Entertainment: $0%$\n",
      "Active lifestyle: $0%$\n",
      "\n",
      "Age group probabilities:\n",
      "- 0-11 years old: $0%$\n",
      "- 12-17 years old: $0%$\n",
      "- 18-21 years old: $20%$\n",
      "- 22-25 years old: $40%$\n",
      "- 26-30 years old: $30%$\n",
      "- 31-40 years old: $10%$\n",
      "- 41-50 years old: $0%$\n",
      "- 51-60 years old: $0%$\n",
      "- Older than 60 years old: $0%$\n",
      "{'df':                  prob\n",
      "nicotine vaping   1.0\n",
      "cannabis vaping   0.0\n",
      "fashion           0.0\n",
      "entertainment     0.0\n",
      "lifestyle         0.0\n",
      "age_0-11          0.0\n",
      "age_12-17         0.0\n",
      "age_18-21         0.2\n",
      "age_22-25         0.4\n",
      "age_26-30         0.3\n",
      "age_31-40         0.1\n",
      "age_41-50         0.0\n",
      "age_51-60         0.0\n",
      "age_61_plus       0.0, 'extracted_probs': 14}\n",
      "\n",
      "Analyzing vapes_aby_ecigs_1\n",
      "1052 frames read.\n",
      "- Probability of nicotine vaping: $100%$\n",
      "- Probability of cannabis vaping: $10%$\n",
      "- Probability of fashion (modeling): $0%$\n",
      "- Probability of entertainment: $20%$\n",
      "- Probability of active lifestyle (sport): $0%$\n",
      "- Probability of age 0-11: $10%$\n",
      "- Probability of age 12-17: $10%$\n",
      "- Probability of age 18-21: $20%$\n",
      "- Probability of age 22-25: $40%$\n",
      "- Probability of age 26-30: $20%$\n",
      "- Probability of age 31-40: $0%$\n",
      "- Probability of age 41-50: $0%$\n",
      "- Probability of age 51-60: $0%$\n",
      "- Probability of older than 60 years old: $0%$\n",
      "{'df':                  prob\n",
      "nicotine vaping   1.0\n",
      "cannabis vaping   0.1\n",
      "fashion           0.0\n",
      "entertainment     0.2\n",
      "lifestyle         0.0\n",
      "age_0-11          0.1\n",
      "age_12-17         0.1\n",
      "age_18-21         0.2\n",
      "age_22-25         0.4\n",
      "age_26-30         0.2\n",
      "age_31-40         0.0\n",
      "age_41-50         0.0\n",
      "age_51-60         0.0\n",
      "age_61_plus       0.0, 'extracted_probs': 14}\n",
      "\n",
      "Analyzing edripss_ecgis_1\n",
      "690 frames read.\n",
      "- Probability of nicotine vaping: $90%$\n",
      "- Probability of cannabis vaping: $10%$\n",
      "- Probability of fashion (modeling): $0%$\n",
      "- Probability of entertainment: $20%$\n",
      "- Probability of active lifestyle (sport): $0%$\n",
      "- Probability of age 0-11: $0%$\n",
      "- Probability of age 12-17: $0%$\n",
      "- Probability of age 18-21: $20%$\n",
      "- Probability of age 22-25: $50%$\n",
      "- Probability of age 26-30: $30%$\n",
      "- Probability of age 31-40: $0%$\n",
      "- Probability of age 41-50: $0%$\n",
      "- Probability of age 51-60: $0%$\n",
      "- Probability of older than 60 years old: $0%$\n",
      "{'df':                  prob\n",
      "nicotine vaping   0.9\n",
      "cannabis vaping   0.1\n",
      "fashion           0.0\n",
      "entertainment     0.2\n",
      "lifestyle         0.0\n",
      "age_0-11          0.0\n",
      "age_12-17         0.0\n",
      "age_18-21         0.2\n",
      "age_22-25         0.5\n",
      "age_26-30         0.3\n",
      "age_31-40         0.0\n",
      "age_41-50         0.0\n",
      "age_51-60         0.0\n",
      "age_61_plus       0.0, 'extracted_probs': 14}\n",
      "\n",
      "Analyzing vapes_aby_ecigs_2\n",
      "779 frames read.\n",
      "- Probability of nicotine vaping: $90%$\n",
      "- Probability of cannabis vaping: $10%$\n",
      "- Probability of fashion (modeling): $0%$\n",
      "- Probability of entertainment: $80%$\n",
      "- Probability of active lifestyle (sport): $0%$\n",
      "\n",
      "- Probability of age 0-11: $0%$\n",
      "- Probability of age 12-17: $10%$\n",
      "- Probability of age 18-21: $30%$\n",
      "- Probability of age 22-25: $40%$\n",
      "- Probability of age 26-30: $20%$\n",
      "- Probability of age 31-40: $0%$\n",
      "- Probability of age 41-50: $0%$\n",
      "- Probability of age 51-60: $0%$\n",
      "- Probability of older than 60 years old: $0%$\n",
      "{'df':                  prob\n",
      "nicotine vaping   0.9\n",
      "cannabis vaping   0.1\n",
      "fashion           0.0\n",
      "entertainment     0.8\n",
      "lifestyle         0.0\n",
      "age_0-11          0.0\n",
      "age_12-17         0.1\n",
      "age_18-21         0.3\n",
      "age_22-25         0.4\n",
      "age_26-30         0.2\n",
      "age_31-40         0.0\n",
      "age_41-50         0.0\n",
      "age_51-60         0.0\n",
      "age_61_plus       0.0, 'extracted_probs': 14}\n",
      "\n",
      "Analyzing arabella_ecigs_4\n",
      "399 frames read.\n",
      "Nicotine vaping: $90%$\n",
      "Cannabis vaping: $10%$\n",
      "Fashion (modeling): $20%$\n",
      "Entertainment (dj, videogames): $80%$\n",
      "Active lifestyle (sport): $10%$\n",
      "\n",
      "Age group probabilities:\n",
      "0-11 years old: $0%$\n",
      "12-17 years old: $10%$\n",
      "18-21 years old: $30%$\n",
      "22-25 years old: $40%$\n",
      "26-30 years old: $15%$\n",
      "31-40 years old: $5%$\n",
      "41-50 years old: $0%$\n",
      "51-60 years old: $0%$\n",
      "Older than 60 years old: $0%$\n",
      "{'df':                  prob\n",
      "nicotine vaping  0.90\n",
      "cannabis vaping  0.10\n",
      "fashion          0.20\n",
      "entertainment    0.80\n",
      "lifestyle        0.10\n",
      "age_0-11         0.00\n",
      "age_12-17        0.10\n",
      "age_18-21        0.30\n",
      "age_22-25        0.40\n",
      "age_26-30        0.15\n",
      "age_31-40        0.05\n",
      "age_41-50        0.00\n",
      "age_51-60        0.00\n",
      "age_61_plus      0.00, 'extracted_probs': 14}\n",
      "\n",
      "Analyzing arabella_ecigs_2\n",
      "1804 frames read.\n",
      "Nicotine vaping: $100%$\n",
      "Cannabis vaping: $10%$\n",
      "Fashion (modeling): $20%$\n",
      "Entertainment: $10%$\n",
      "Active lifestyle (sport): $10%$\n",
      "\n",
      "Age 0-11: $0%$\n",
      "Age 12-17: $10%$\n",
      "Age 18-21: $30%$\n",
      "Age 22-25: $40%$\n",
      "Age 26-30: $15%$\n",
      "Age 31-40: $5%$\n",
      "Age 41-50: $0%$\n",
      "Age 51-60: $0%$\n",
      "Older than 60: $0%$\n",
      "{'df':                  prob\n",
      "nicotine vaping  1.00\n",
      "cannabis vaping  0.10\n",
      "fashion          0.20\n",
      "entertainment    0.10\n",
      "lifestyle        0.10\n",
      "age_0-11         0.00\n",
      "age_12-17        0.10\n",
      "age_18-21        0.30\n",
      "age_22-25        0.40\n",
      "age_26-30        0.15\n",
      "age_31-40        0.05\n",
      "age_41-50        0.00\n",
      "age_51-60        0.00\n",
      "age_61_plus      0.00, 'extracted_probs': 14}\n",
      "\n",
      "Analyzing drewdirps_ecigs_1\n",
      "350 frames read.\n",
      "- Probability of nicotine vaping: $90%$\n",
      "- Probability of cannabis vaping: $10%$\n",
      "- Probability of fashion (modeling): $0%$\n",
      "- Probability of entertainment: $0%$\n",
      "- Probability of active lifestyle (sport): $0%$\n",
      "- Probability of age 0-11: $0%$\n",
      "- Probability of age 12-17: $0%$\n",
      "- Probability of age 18-21: $10%$\n",
      "- Probability of age 22-25: $30%$\n",
      "- Probability of age 26-30: $40%$\n",
      "- Probability of age 31-40: $20%$\n",
      "- Probability of age 41-50: $0%$\n",
      "- Probability of age 51-60: $0%$\n",
      "- Probability of older than 60 years old: $0%$\n",
      "{'df':                  prob\n",
      "nicotine vaping   0.9\n",
      "cannabis vaping   0.1\n",
      "fashion           0.0\n",
      "entertainment     0.0\n",
      "lifestyle         0.0\n",
      "age_0-11          0.0\n",
      "age_12-17         0.0\n",
      "age_18-21         0.1\n",
      "age_22-25         0.3\n",
      "age_26-30         0.4\n",
      "age_31-40         0.2\n",
      "age_41-50         0.0\n",
      "age_51-60         0.0\n",
      "age_61_plus       0.0, 'extracted_probs': 14}\n",
      "\n",
      "Analyzing arabella_ecigs_5\n",
      "156 frames read.\n",
      "- Nicotine vaping: $80%$\n",
      "- Cannabis vaping: $10%$\n",
      "- Fashion (modeling): $30%$\n",
      "- Entertainment: $20%$\n",
      "- Active lifestyle (sport): $10%$\n",
      "\n",
      "- Age 0-11: $0%$\n",
      "- Age 12-17: $0%$\n",
      "- Age 18-21: $20%$\n",
      "- Age 22-25: $50%$\n",
      "- Age 26-30: $20%$\n",
      "- Age 31-40: $10%$\n",
      "- Age 41-50: $0%$\n",
      "- Age 51-60: $0%$\n",
      "- Older than 60: $0%$\n",
      "{'df':                  prob\n",
      "nicotine vaping   0.8\n",
      "cannabis vaping   0.1\n",
      "fashion           0.3\n",
      "entertainment     0.2\n",
      "lifestyle         0.1\n",
      "age_0-11          0.0\n",
      "age_12-17         0.0\n",
      "age_18-21         0.2\n",
      "age_22-25         0.5\n",
      "age_26-30         0.2\n",
      "age_31-40         0.1\n",
      "age_41-50         0.0\n",
      "age_51-60         0.0\n",
      "age_61_plus       0.0, 'extracted_probs': 14}\n",
      "\n",
      "Analyzing calitrickzz_ecigs_2\n",
      "825 frames read.\n",
      "Nicotine vaping: $100%$\n",
      "Cannabis vaping: $0%$\n",
      "Fashion (modeling): $0%$\n",
      "Entertainment: $0%$\n",
      "Active lifestyle (sport): $0%$\n",
      "\n",
      "Age group probabilities:\n",
      "0-11 years old: $0%$\n",
      "12-17 years old: $0%$\n",
      "18-21 years old: $20%$\n",
      "22-25 years old: $50%$\n",
      "26-30 years old: $20%$\n",
      "31-40 years old: $10%$\n",
      "41-50 years old: $0%$\n",
      "51-60 years old: $0%$\n",
      "Older than 60 years old: $0%$\n",
      "{'df':                  prob\n",
      "nicotine vaping   1.0\n",
      "cannabis vaping   0.0\n",
      "fashion           0.0\n",
      "entertainment     0.0\n",
      "lifestyle         0.0\n",
      "age_0-11          0.0\n",
      "age_12-17         0.0\n",
      "age_18-21         0.2\n",
      "age_22-25         0.5\n",
      "age_26-30         0.2\n",
      "age_31-40         0.1\n",
      "age_41-50         0.0\n",
      "age_51-60         0.0\n",
      "age_61_plus       0.0, 'extracted_probs': 14}\n",
      "\n",
      "Analyzing alex_ecigs_1\n",
      "295 frames read.\n",
      "Nicotine vaping: $80%$\n",
      "Cannabis vaping: $70%$\n",
      "Fashion (modeling): $20%$\n",
      "Entertainment: $30%$\n",
      "Active lifestyle: $10%$\n",
      "\n",
      "Age 0-11: $0%$\n",
      "Age 12-17: $10%$\n",
      "Age 18-21: $30%$\n",
      "Age 22-25: $40%$\n",
      "Age 26-30: $15%$\n",
      "Age 31-40: $5%$\n",
      "Age 41-50: $0%$\n",
      "Age 51-60: $0%$\n",
      "Older than 60: $0%$\n",
      "{'df':                  prob\n",
      "nicotine vaping  0.80\n",
      "cannabis vaping  0.70\n",
      "fashion          0.20\n",
      "entertainment    0.30\n",
      "lifestyle        0.10\n",
      "age_0-11         0.00\n",
      "age_12-17        0.10\n",
      "age_18-21        0.30\n",
      "age_22-25        0.40\n",
      "age_26-30        0.15\n",
      "age_31-40        0.05\n",
      "age_41-50        0.00\n",
      "age_51-60        0.00\n",
      "age_61_plus      0.00, 'extracted_probs': 14}\n",
      "\n",
      "Analyzing fabio_ecigs_1\n",
      "430 frames read.\n",
      "- Probability of nicotine vaping: $0%$\n",
      "- Probability of cannabis vaping: $0%$\n",
      "- Probability of fashion (modeling): $10%$\n",
      "- Probability of entertainment: $10%$\n",
      "- Probability of active lifestyle (sport): $10%$\n",
      "- Probability of age 0-11: $0%$\n",
      "- Probability of age 12-17: $0%$\n",
      "- Probability of age 18-21: $10%$\n",
      "- Probability of age 22-25: $30%$\n",
      "- Probability of age 26-30: $40%$\n",
      "- Probability of age 31-40: $20%$\n",
      "- Probability of age 41-50: $0%$\n",
      "- Probability of age 51-60: $0%$\n",
      "- Probability of older than 60 years old: $0%$\n",
      "{'df':                  prob\n",
      "nicotine vaping   0.0\n",
      "cannabis vaping   0.0\n",
      "fashion           0.1\n",
      "entertainment     0.1\n",
      "lifestyle         0.1\n",
      "age_0-11          0.0\n",
      "age_12-17         0.0\n",
      "age_18-21         0.1\n",
      "age_22-25         0.3\n",
      "age_26-30         0.4\n",
      "age_31-40         0.2\n",
      "age_41-50         0.0\n",
      "age_51-60         0.0\n",
      "age_61_plus       0.0, 'extracted_probs': 14}\n",
      "\n",
      "Analyzing drewdirps_ecigs_2\n",
      "751 frames read.\n",
      "Nicotine vaping: $100%$\n",
      "Cannabis vaping: $10%$\n",
      "Fashion (modeling): $0%$\n",
      "Entertainment: $20%$\n",
      "Active lifestyle (sport): $0%$\n",
      "\n",
      "Age group probabilities:\n",
      "- 0-11 years old: $0%$\n",
      "- 12-17 years old: $0%$\n",
      "- 18-21 years old: $10%$\n",
      "- 22-25 years old: $20%$\n",
      "- 26-30 years old: $30%$\n",
      "- 31-40 years old: $30%$\n",
      "- 41-50 years old: $10%$\n",
      "- 51-60 years old: $0%$\n",
      "- Older than 60 years old: $0%$\n",
      "{'df':                  prob\n",
      "nicotine vaping   1.0\n",
      "cannabis vaping   0.1\n",
      "fashion           0.0\n",
      "entertainment     0.2\n",
      "lifestyle         0.0\n",
      "age_0-11          0.0\n",
      "age_12-17         0.0\n",
      "age_18-21         0.1\n",
      "age_22-25         0.2\n",
      "age_26-30         0.3\n",
      "age_31-40         0.3\n",
      "age_41-50         0.1\n",
      "age_51-60         0.0\n",
      "age_61_plus       0.0, 'extracted_probs': 14}\n",
      "\n",
      "Analyzing arabella_ecigs_1\n",
      "520 frames read.\n",
      "Nicotine vaping: $100%$\n",
      "Cannabis vaping: $0%$\n",
      "Fashion (modeling): $10%$\n",
      "Entertainment: $10%$\n",
      "Active lifestyle (sport): $0%$\n",
      "\n",
      "Age probability:\n",
      "- 0-11 years old: $0%$\n",
      "- 12-17 years old: $0%$\n",
      "- 18-21 years old: $20%$\n",
      "- 22-25 years old: $50%$\n",
      "- 26-30 years old: $20%$\n",
      "- 31-40 years old: $10%$\n",
      "- 41-50 years old: $0%$\n",
      "- 51-60 years old: $0%$\n",
      "- Older than 60 years old: $0%$\n",
      "{'df':                  prob\n",
      "nicotine vaping   1.0\n",
      "cannabis vaping   0.0\n",
      "fashion           0.1\n",
      "entertainment     0.1\n",
      "lifestyle         0.0\n",
      "age_0-11          0.0\n",
      "age_12-17         0.0\n",
      "age_18-21         0.2\n",
      "age_22-25         0.5\n",
      "age_26-30         0.2\n",
      "age_31-40         0.1\n",
      "age_41-50         0.0\n",
      "age_51-60         0.0\n",
      "age_61_plus       0.0, 'extracted_probs': 14}\n",
      "\n",
      "Analyzing zanxlyfe_ecigs_1\n",
      "2554 frames read.\n",
      "Nicotine vaping: $90%$\n",
      "Cannabis vaping: $10%$\n",
      "Fashion (modeling): $0%$\n",
      "Entertainment: $50%$\n",
      "Active lifestyle (sport): $0%$\n",
      "\n",
      "Age probability:\n",
      "0-11 years old: $0%$\n",
      "12-17 years old: $10%$\n",
      "18-21 years old: $30%$\n",
      "22-25 years old: $40%$\n",
      "26-30 years old: $15%$\n",
      "31-40 years old: $5%$\n",
      "41-50 years old: $0%$\n",
      "51-60 years old: $0%$\n",
      "Older than 60 years old: $0%$\n",
      "{'df':                  prob\n",
      "nicotine vaping  0.90\n",
      "cannabis vaping  0.10\n",
      "fashion          0.00\n",
      "entertainment    0.50\n",
      "lifestyle        0.00\n",
      "age_0-11         0.00\n",
      "age_12-17        0.10\n",
      "age_18-21        0.30\n",
      "age_22-25        0.40\n",
      "age_26-30        0.15\n",
      "age_31-40        0.05\n",
      "age_41-50        0.00\n",
      "age_51-60        0.00\n",
      "age_61_plus      0.00, 'extracted_probs': 14}\n",
      "\n",
      "Analyzing arabella_ecigs_3\n",
      "655 frames read.\n",
      "- Probability of nicotine vaping: $100%$\n",
      "- Probability of cannabis vaping: $10%$\n",
      "- Probability of fashion (modeling): $0%$\n",
      "- Probability of entertainment: $10%$\n",
      "- Probability of active lifestyle (sport): $0%$\n",
      "\n",
      "- Probability that the person in this video is between the age of 0-11: $0%$\n",
      "- Probability that the person in this video is between the age of 12-17: $80%$\n",
      "- Probability that the person in this video is between the age of 18-21: $20%$\n",
      "- Probability that the person in this video is between the age of 22-25: $0%$\n",
      "- Probability that the person in this video is between the age of 26-30: $0%$\n",
      "- Probability that the person in this video is between the age of 31-40: $0%$\n",
      "- Probability that the person in this video is between the age of 41\n",
      "{'df':                  prob\n",
      "nicotine vaping   1.0\n",
      "cannabis vaping   0.1\n",
      "fashion           0.0\n",
      "entertainment     0.1\n",
      "lifestyle         0.0\n",
      "age_0-11          0.0\n",
      "age_12-17         0.8\n",
      "age_18-21         0.2\n",
      "age_22-25         0.0\n",
      "age_26-30         0.0\n",
      "age_31-40         0.0\n",
      "age_41-50        None\n",
      "age_51-60        None\n",
      "age_61_plus      None, 'extracted_probs': 11}\n",
      "\n",
      "Analyzing calitrickzz_ecigs_1\n",
      "503 frames read.\n",
      "Nicotine vaping: $100%$\n",
      "Cannabis vaping: $10%$\n",
      "Fashion (modeling): $0%$\n",
      "Entertainment (dj, videogames): $0%$\n",
      "Active lifestyle (sport): $0%$\n",
      "\n",
      "Age group probabilities:\n",
      "0-11 years old: $0%$\n",
      "12-17 years old: $0%$\n",
      "18-21 years old: $20%$\n",
      "22-25 years old: $50%$\n",
      "26-30 years old: $20%$\n",
      "31-40 years old: $10%$\n",
      "41-50 years old: $0%$\n",
      "51-60 years old: $0%$\n",
      "Older than 60 years old: $0%$\n",
      "{'df':                  prob\n",
      "nicotine vaping   1.0\n",
      "cannabis vaping   0.1\n",
      "fashion           0.0\n",
      "entertainment     0.0\n",
      "lifestyle         0.0\n",
      "age_0-11          0.0\n",
      "age_12-17         0.0\n",
      "age_18-21         0.2\n",
      "age_22-25         0.5\n",
      "age_26-30         0.2\n",
      "age_31-40         0.1\n",
      "age_41-50         0.0\n",
      "age_51-60         0.0\n",
      "age_61_plus       0.0, 'extracted_probs': 14}\n",
      "\n",
      "Analyzing zanxlyfe_ecigs_2\n",
      "1669 frames read.\n",
      "- Nicotine vaping: $90%$\n",
      "- Cannabis vaping: $10%$\n",
      "- Fashion (modeling): $0%$\n",
      "- Entertainment: $20%$\n",
      "- Active lifestyle (sport): $0%$\n",
      "\n",
      "- Age 0-11: $0%$\n",
      "- Age 12-17: $10%$\n",
      "- Age 18-21: $30%$\n",
      "- Age 22-25: $40%$\n",
      "- Age 26-30: $15%$\n",
      "- Age 31-40: $5%$\n",
      "- Age 41-50: $0%$\n",
      "- Age 51-60: $0%$\n",
      "- Older than 60: $0%$\n",
      "{'df':                  prob\n",
      "nicotine vaping  0.90\n",
      "cannabis vaping  0.10\n",
      "fashion          0.00\n",
      "entertainment    0.20\n",
      "lifestyle        0.00\n",
      "age_0-11         0.00\n",
      "age_12-17        0.10\n",
      "age_18-21        0.30\n",
      "age_22-25        0.40\n",
      "age_26-30        0.15\n",
      "age_31-40        0.05\n",
      "age_41-50        0.00\n",
      "age_51-60        0.00\n",
      "age_61_plus      0.00, 'extracted_probs': 14}\n",
      "\n",
      "Analyzing zanxlyfe_tech_1\n",
      "2701 frames read.\n",
      "Nicotine vaping: $0%$\n",
      "Cannabis vaping: $0%$\n",
      "Fashion (modeling): $0%$\n",
      "Entertainment: $100%$\n",
      "Active lifestyle: $0%$\n",
      "\n",
      "Age group probabilities:\n",
      "0-11 years old: $0%$\n",
      "12-17 years old: $10%$\n",
      "18-21 years old: $30%$\n",
      "22-25 years old: $40%$\n",
      "26-30 years old: $15%$\n",
      "31-40 years old: $5%$\n",
      "41-50 years old: $0%$\n",
      "51-60 years old: $0%$\n",
      "Older than 60 years old: $0%$\n",
      "{'df':                  prob\n",
      "nicotine vaping  0.00\n",
      "cannabis vaping  0.00\n",
      "fashion          0.00\n",
      "entertainment    1.00\n",
      "lifestyle        0.00\n",
      "age_0-11         0.00\n",
      "age_12-17        0.10\n",
      "age_18-21        0.30\n",
      "age_22-25        0.40\n",
      "age_26-30        0.15\n",
      "age_31-40        0.05\n",
      "age_41-50        0.00\n",
      "age_51-60        0.00\n",
      "age_61_plus      0.00, 'extracted_probs': 14}\n",
      "\n",
      "Analyzing arabella_tech_2\n",
      "1719 frames read.\n",
      "Nicotine vaping: $0%$\n",
      "Cannabis vaping: $0%$\n",
      "Fashion (modeling): $0%$\n",
      "Entertainment (DJ, videogames): $100%$\n",
      "Active lifestyle (sport): $0%$\n",
      "\n",
      "Age 0-11: $0%$\n",
      "Age 12-17: $0%$\n",
      "Age 18-21: $20%$\n",
      "Age 22-25: $50%$\n",
      "Age 26-30: $20%$\n",
      "Age 31-40: $10%$\n",
      "Age 41-50: $0%$\n",
      "Age 51-60: $0%$\n",
      "Older than 60: $0%$\n",
      "{'df':                  prob\n",
      "nicotine vaping   0.0\n",
      "cannabis vaping   0.0\n",
      "fashion           0.0\n",
      "entertainment     1.0\n",
      "lifestyle         0.0\n",
      "age_0-11          0.0\n",
      "age_12-17         0.0\n",
      "age_18-21         0.2\n",
      "age_22-25         0.5\n",
      "age_26-30         0.2\n",
      "age_31-40         0.1\n",
      "age_41-50         0.0\n",
      "age_51-60         0.0\n",
      "age_61_plus       0.0, 'extracted_probs': 14}\n",
      "\n",
      "Analyzing zanxlyfe_tech_5\n",
      "1672 frames read.\n",
      "Nicotine vaping: $0%$\n",
      "Cannabis vaping: $0%$\n",
      "Fashion (modeling): $0%$\n",
      "Entertainment: $100%$\n",
      "Active lifestyle: $0%$\n",
      "\n",
      "Age 0-11: $0%$\n",
      "Age 12-17: $0%$\n",
      "Age 18-21: $20%$\n",
      "Age 22-25: $50%$\n",
      "Age 26-30: $20%$\n",
      "Age 31-40: $10%$\n",
      "Age 41-50: $0%$\n",
      "Age 51-60: $0%$\n",
      "Older than 60: $0%$\n",
      "{'df':                  prob\n",
      "nicotine vaping   0.0\n",
      "cannabis vaping   0.0\n",
      "fashion           0.0\n",
      "entertainment     1.0\n",
      "lifestyle         0.0\n",
      "age_0-11          0.0\n",
      "age_12-17         0.0\n",
      "age_18-21         0.2\n",
      "age_22-25         0.5\n",
      "age_26-30         0.2\n",
      "age_31-40         0.1\n",
      "age_41-50         0.0\n",
      "age_51-60         0.0\n",
      "age_61_plus       0.0, 'extracted_probs': 14}\n",
      "\n",
      "Analyzing zanxlyfe_tech_6\n",
      "2701 frames read.\n",
      "Nicotine vaping: $0%$\n",
      "Cannabis vaping: $0%$\n",
      "Fashion (modeling): $0%$\n",
      "Entertainment: $100%$\n",
      "Active lifestyle (sport): $0%$\n",
      "\n",
      "Age group probabilities:\n",
      "- 0-11 years old: $0%$\n",
      "- 12-17 years old: $10%$\n",
      "- 18-21 years old: $30%$\n",
      "- 22-25 years old: $40%$\n",
      "- 26-30 years old: $15%$\n",
      "- 31-40 years old: $5%$\n",
      "- 41-50 years old: $0%$\n",
      "- 51-60 years old: $0%$\n",
      "- Older than 60 years old: $0%$\n",
      "{'df':                  prob\n",
      "nicotine vaping  0.00\n",
      "cannabis vaping  0.00\n",
      "fashion          0.00\n",
      "entertainment    1.00\n",
      "lifestyle        0.00\n",
      "age_0-11         0.00\n",
      "age_12-17        0.10\n",
      "age_18-21        0.30\n",
      "age_22-25        0.40\n",
      "age_26-30        0.15\n",
      "age_31-40        0.05\n",
      "age_41-50        0.00\n",
      "age_51-60        0.00\n",
      "age_61_plus      0.00, 'extracted_probs': 14}\n",
      "\n",
      "Analyzing zanxlyfe_tech_10\n",
      "1087 frames read.\n",
      "Nicotine vaping: $0%$\n",
      "Cannabis vaping: $0%$\n",
      "Fashion (modeling): $0%$\n",
      "Entertainment (DJ, videogames): $100%$\n",
      "Active lifestyle (sport): $0%$\n",
      "\n",
      "Age group probabilities:\n",
      "- 0-11 years old: $0%$\n",
      "- 12-17 years old: $0%$\n",
      "- 18-21 years old: $20%$\n",
      "- 22-25 years old: $60%$\n",
      "- 26-30 years old: $20%$\n",
      "- 31-40 years old: $0%$\n",
      "- 41-50 years old: $0%$\n",
      "- 51-60 years old: $0%$\n",
      "- Older than 60 years old: $0%$\n",
      "{'df':                  prob\n",
      "nicotine vaping   0.0\n",
      "cannabis vaping   0.0\n",
      "fashion           0.0\n",
      "entertainment     1.0\n",
      "lifestyle         0.0\n",
      "age_0-11          0.0\n",
      "age_12-17         0.0\n",
      "age_18-21         0.2\n",
      "age_22-25         0.6\n",
      "age_26-30         0.2\n",
      "age_31-40         0.0\n",
      "age_41-50         0.0\n",
      "age_51-60         0.0\n",
      "age_61_plus       0.0, 'extracted_probs': 14}\n",
      "\n",
      "Analyzing zanxlyfe_tech_4\n",
      "1712 frames read.\n",
      "Nicotine vaping: $0%$\n",
      "Cannabis vaping: $0%$\n",
      "Fashion (modeling): $0%$\n",
      "Entertainment: $100%$\n",
      "Active lifestyle (sport): $0%$\n",
      "\n",
      "Age group probabilities:\n",
      "- 0-11 years old: $0%$\n",
      "- 12-17 years old: $0%$\n",
      "- 18-21 years old: $20%$\n",
      "- 22-25 years old: $50%$\n",
      "- 26-30 years old: $30%$\n",
      "- 31-40 years old: $0%$\n",
      "- 41-50 years old: $0%$\n",
      "- 51-60 years old: $0%$\n",
      "- Older than 60 years old: $0%$\n",
      "{'df':                  prob\n",
      "nicotine vaping   0.0\n",
      "cannabis vaping   0.0\n",
      "fashion           0.0\n",
      "entertainment     1.0\n",
      "lifestyle         0.0\n",
      "age_0-11          0.0\n",
      "age_12-17         0.0\n",
      "age_18-21         0.2\n",
      "age_22-25         0.5\n",
      "age_26-30         0.3\n",
      "age_31-40         0.0\n",
      "age_41-50         0.0\n",
      "age_51-60         0.0\n",
      "age_61_plus       0.0, 'extracted_probs': 14}\n",
      "\n",
      "Analyzing arabella_tech_4\n",
      "1609 frames read.\n",
      "- Probability of nicotine vaping: $10%$\n",
      "- Probability of cannabis vaping: $10%$\n",
      "- Probability of fashion (modeling): $10%$\n",
      "- Probability of entertainment (e.g., DJ, video games): $100%$\n",
      "- Probability of active lifestyle (e.g., sport): $10%$\n",
      "\n",
      "- Probability that the person in this video is between the age of 0-11: $0%$\n",
      "- Probability that the person in this video is between the age of 12-17: $10%$\n",
      "- Probability that the person in this video is between the age of 18-21: $40%$\n",
      "- Probability that the person in this video is between the age of 22-25: $40%$\n",
      "- Probability that the person in this video is between the age of 26-30: $10%$\n",
      "- Probability that the person in this video is between the age of 31-40: $0%$\n",
      "- Probability that the\n",
      "{'df':                  prob\n",
      "nicotine vaping   0.1\n",
      "cannabis vaping   0.1\n",
      "fashion           0.1\n",
      "entertainment     1.0\n",
      "lifestyle         0.1\n",
      "age_0-11          0.0\n",
      "age_12-17         0.1\n",
      "age_18-21         0.4\n",
      "age_22-25         0.4\n",
      "age_26-30         0.1\n",
      "age_31-40         0.0\n",
      "age_41-50        None\n",
      "age_51-60        None\n",
      "age_61_plus      None, 'extracted_probs': 11}\n",
      "\n",
      "Analyzing zanxlyfe_tech_8\n",
      "1368 frames read.\n",
      "- Nicotine vaping: $0%$\n",
      "- Cannabis vaping: $0%$\n",
      "- Fashion (modeling): $0%$\n",
      "- Entertainment (DJ, video games): $100%$\n",
      "- Active lifestyle (sport): $0%$\n",
      "\n",
      "- Age 0-11: $0%$\n",
      "- Age 12-17: $0%$\n",
      "- Age 18-21: $20%$\n",
      "- Age 22-25: $50%$\n",
      "- Age 26-30: $20%$\n",
      "- Age 31-40: $10%$\n",
      "- Age 41-50: $0%$\n",
      "- Age 51-60: $0%$\n",
      "- Older than 60: $0%$\n",
      "{'df':                  prob\n",
      "nicotine vaping   0.0\n",
      "cannabis vaping   0.0\n",
      "fashion           0.0\n",
      "entertainment     1.0\n",
      "lifestyle         0.0\n",
      "age_0-11          0.0\n",
      "age_12-17         0.0\n",
      "age_18-21         0.2\n",
      "age_22-25         0.5\n",
      "age_26-30         0.2\n",
      "age_31-40         0.1\n",
      "age_41-50         0.0\n",
      "age_51-60         0.0\n",
      "age_61_plus       0.0, 'extracted_probs': 14}\n",
      "\n",
      "Analyzing arabella_tech_7\n",
      "2323 frames read.\n",
      "Nicotine vaping: $0%$\n",
      "Cannabis vaping: $0%$\n",
      "Fashion (modeling): $0%$\n",
      "Entertainment (DJ, videogames): $100%$\n",
      "Active lifestyle (sport): $0%$\n",
      "\n",
      "Age group probabilities:\n",
      "0-11 years old: $0%$\n",
      "12-17 years old: $0%$\n",
      "18-21 years old: $20%$\n",
      "22-25 years old: $50%$\n",
      "26-30 years old: $20%$\n",
      "31-40 years old: $10%$\n",
      "41-50 years old: $0%$\n",
      "51-60 years old: $0%$\n",
      "Older than 60 years old: $0%$\n",
      "{'df':                  prob\n",
      "nicotine vaping   0.0\n",
      "cannabis vaping   0.0\n",
      "fashion           0.0\n",
      "entertainment     1.0\n",
      "lifestyle         0.0\n",
      "age_0-11          0.0\n",
      "age_12-17         0.0\n",
      "age_18-21         0.2\n",
      "age_22-25         0.5\n",
      "age_26-30         0.2\n",
      "age_31-40         0.1\n",
      "age_41-50         0.0\n",
      "age_51-60         0.0\n",
      "age_61_plus       0.0, 'extracted_probs': 14}\n",
      "\n",
      "Analyzing arabella_tech_10\n",
      "1165 frames read.\n",
      "Nicotine vaping: $0%$\n",
      "Cannabis vaping: $0%$\n",
      "Fashion (modeling): $0%$\n",
      "Entertainment (DJ, videogames): $100%$\n",
      "Active lifestyle (sport): $0%$\n",
      "\n",
      "Age 0-11: $0%$\n",
      "Age 12-17: $0%$\n",
      "Age 18-21: $10%$\n",
      "Age 22-25: $40%$\n",
      "Age 26-30: $40%$\n",
      "Age 31-40: $10%$\n",
      "Age 41-50: $0%$\n",
      "Age 51-60: $0%$\n",
      "Older than 60: $0%$\n",
      "{'df':                  prob\n",
      "nicotine vaping   0.0\n",
      "cannabis vaping   0.0\n",
      "fashion           0.0\n",
      "entertainment     1.0\n",
      "lifestyle         0.0\n",
      "age_0-11          0.0\n",
      "age_12-17         0.0\n",
      "age_18-21         0.1\n",
      "age_22-25         0.4\n",
      "age_26-30         0.4\n",
      "age_31-40         0.1\n",
      "age_41-50         0.0\n",
      "age_51-60         0.0\n",
      "age_61_plus       0.0, 'extracted_probs': 14}\n",
      "\n",
      "Analyzing arabella_tech_1\n",
      "1663 frames read.\n",
      "Nicotine vaping: $0%$\n",
      "Cannabis vaping: $0%$\n",
      "Fashion (modeling): $0%$\n",
      "Entertainment (dj, videogames): $100%$\n",
      "Active lifestyle (sport): $0%$\n",
      "\n",
      "Age group probabilities:\n",
      "0-11 years old: $0%$\n",
      "12-17 years old: $0%$\n",
      "18-21 years old: $20%$\n",
      "22-25 years old: $50%$\n",
      "26-30 years old: $20%$\n",
      "31-40 years old: $10%$\n",
      "41-50 years old: $0%$\n",
      "51-60 years old: $0%$\n",
      "Older than 60 years old: $0%$\n",
      "{'df':                  prob\n",
      "nicotine vaping   0.0\n",
      "cannabis vaping   0.0\n",
      "fashion           0.0\n",
      "entertainment     1.0\n",
      "lifestyle         0.0\n",
      "age_0-11          0.0\n",
      "age_12-17         0.0\n",
      "age_18-21         0.2\n",
      "age_22-25         0.5\n",
      "age_26-30         0.2\n",
      "age_31-40         0.1\n",
      "age_41-50         0.0\n",
      "age_51-60         0.0\n",
      "age_61_plus       0.0, 'extracted_probs': 14}\n",
      "\n",
      "Analyzing zanxlyfe_tech_3\n",
      "1604 frames read.\n",
      "- Probability of nicotine vaping: $0%$\n",
      "- Probability of cannabis vaping: $0%$\n",
      "- Probability of fashion (modeling): $0%$\n",
      "- Probability of entertainment (e.g., DJ, video games): $100%$\n",
      "- Probability of active lifestyle (e.g., sport): $0%$\n",
      "\n",
      "- Probability of age 0-11: $0%$\n",
      "- Probability of age 12-17: $0%$\n",
      "- Probability of age 18-21: $20%$\n",
      "- Probability of age 22-25: $50%$\n",
      "- Probability of age 26-30: $20%$\n",
      "- Probability of age 31-40: $10%$\n",
      "- Probability of age 41-50: $0%$\n",
      "- Probability of age 51-60: $0%$\n",
      "- Probability of older than 60 years old: $0%$\n",
      "{'df':                  prob\n",
      "nicotine vaping   0.0\n",
      "cannabis vaping   0.0\n",
      "fashion           0.0\n",
      "entertainment     1.0\n",
      "lifestyle         0.0\n",
      "age_0-11          0.0\n",
      "age_12-17         0.0\n",
      "age_18-21         0.2\n",
      "age_22-25         0.5\n",
      "age_26-30         0.2\n",
      "age_31-40         0.1\n",
      "age_41-50         0.0\n",
      "age_51-60         0.0\n",
      "age_61_plus       0.0, 'extracted_probs': 14}\n",
      "\n",
      "Analyzing zanxlyfe_tech_7\n",
      "1509 frames read.\n",
      "Nicotine vaping: $90%$\n",
      "Cannabis vaping: $10%$\n",
      "Fashion (modeling): $0%$\n",
      "Entertainment (DJ, video games): $100%$\n",
      "Active lifestyle (sport): $0%$\n",
      "\n",
      "Age group probabilities:\n",
      "0-11 years old: $0%$\n",
      "12-17 years old: $10%$\n",
      "18-21 years old: $30%$\n",
      "22-25 years old: $40%$\n",
      "26-30 years old: $15%$\n",
      "31-40 years old: $5%$\n",
      "41-50 years old: $0%$\n",
      "51-60 years old: $0%$\n",
      "Older than 60 years old: $0%$\n",
      "{'df':                  prob\n",
      "nicotine vaping  0.90\n",
      "cannabis vaping  0.10\n",
      "fashion          0.00\n",
      "entertainment    1.00\n",
      "lifestyle        0.00\n",
      "age_0-11         0.00\n",
      "age_12-17        0.10\n",
      "age_18-21        0.30\n",
      "age_22-25        0.40\n",
      "age_26-30        0.15\n",
      "age_31-40        0.05\n",
      "age_41-50        0.00\n",
      "age_51-60        0.00\n",
      "age_61_plus      0.00, 'extracted_probs': 14}\n",
      "\n",
      "Analyzing zanxlyfe_tech_2\n",
      "2111 frames read.\n",
      "Nicotine vaping: $0%$\n",
      "Cannabis vaping: $0%$\n",
      "Fashion (modeling): $0%$\n",
      "Entertainment: $100%$\n",
      "Active lifestyle (sport): $0%$\n",
      "\n",
      "Age group probabilities:\n",
      "- 0-11 years old: $0%$\n",
      "- 12-17 years old: $10%$\n",
      "- 18-21 years old: $30%$\n",
      "- 22-25 years old: $40%$\n",
      "- 26-30 years old: $15%$\n",
      "- 31-40 years old: $5%$\n",
      "- 41-50 years old: $0%$\n",
      "- 51-60 years old: $0%$\n",
      "- Older than 60 years old: $0%$\n",
      "{'df':                  prob\n",
      "nicotine vaping  0.00\n",
      "cannabis vaping  0.00\n",
      "fashion          0.00\n",
      "entertainment    1.00\n",
      "lifestyle        0.00\n",
      "age_0-11         0.00\n",
      "age_12-17        0.10\n",
      "age_18-21        0.30\n",
      "age_22-25        0.40\n",
      "age_26-30        0.15\n",
      "age_31-40        0.05\n",
      "age_41-50        0.00\n",
      "age_51-60        0.00\n",
      "age_61_plus      0.00, 'extracted_probs': 14}\n",
      "\n",
      "Analyzing arabella_tech_5\n",
      "1275 frames read.\n",
      "Nicotine vaping: $0%$\n",
      "Cannabis vaping: $0%$\n",
      "Fashion (modeling): $0%$\n",
      "Entertainment (DJ, videogames): $100%$\n",
      "Active lifestyle (sport): $0%$\n",
      "\n",
      "Age group probabilities:\n",
      "0-11 years old: $0%$\n",
      "12-17 years old: $0%$\n",
      "18-21 years old: $20%$\n",
      "22-25 years old: $50%$\n",
      "26-30 years old: $20%$\n",
      "31-40 years old: $10%$\n",
      "41-50 years old: $0%$\n",
      "51-60 years old: $0%$\n",
      "Older than 60 years old: $0%$\n",
      "{'df':                  prob\n",
      "nicotine vaping   0.0\n",
      "cannabis vaping   0.0\n",
      "fashion           0.0\n",
      "entertainment     1.0\n",
      "lifestyle         0.0\n",
      "age_0-11          0.0\n",
      "age_12-17         0.0\n",
      "age_18-21         0.2\n",
      "age_22-25         0.5\n",
      "age_26-30         0.2\n",
      "age_31-40         0.1\n",
      "age_41-50         0.0\n",
      "age_51-60         0.0\n",
      "age_61_plus       0.0, 'extracted_probs': 14}\n",
      "\n",
      "Analyzing arabella_tech_6\n",
      "1073 frames read.\n",
      "- Probability of nicotine vaping: $0%$\n",
      "- Probability of cannabis vaping: $0%$\n",
      "- Probability of fashion (modeling): $20%$\n",
      "- Probability of entertainment (DJ, video games): $100%$\n",
      "- Probability of active lifestyle (sport): $0%$\n",
      "\n",
      "- Probability that the person is between the age of 0-11: $0%$\n",
      "- Probability that the person is between the age of 12-17: $0%$\n",
      "- Probability that the person is between the age of 18-21: $20%$\n",
      "- Probability that the person is between the age of 22-25: $60%$\n",
      "- Probability that the person is between the age of 26-30: $20%$\n",
      "- Probability that the person is between the age of 31-40: $0%$\n",
      "- Probability that the person is between the age of 41-50: $0%$\n",
      "- Probability that the person is between the age\n",
      "{'df':                  prob\n",
      "nicotine vaping   0.0\n",
      "cannabis vaping   0.0\n",
      "fashion           0.2\n",
      "entertainment     1.0\n",
      "lifestyle         0.0\n",
      "age_0-11          0.0\n",
      "age_12-17         0.0\n",
      "age_18-21         0.2\n",
      "age_22-25         0.6\n",
      "age_26-30         0.2\n",
      "age_31-40         0.0\n",
      "age_41-50         0.0\n",
      "age_51-60        None\n",
      "age_61_plus      None, 'extracted_probs': 12}\n",
      "\n",
      "Analyzing zanxlyfe_tech_9\n",
      "896 frames read.\n",
      "Nicotine vaping: $0%$\n",
      "Cannabis vaping: $0%$\n",
      "Fashion (modeling): $0%$\n",
      "Entertainment (DJ, videogames): $100%$\n",
      "Active lifestyle (sport): $0%$\n",
      "\n",
      "Age range probabilities:\n",
      "- 0-11 years old: $0%$\n",
      "- 12-17 years old: $0%$\n",
      "- 18-21 years old: $20%$\n",
      "- 22-25 years old: $50%$\n",
      "- 26-30 years old: $30%$\n",
      "- 31-40 years old: $0%$\n",
      "- 41-50 years old: $0%$\n",
      "- 51-60 years old: $0%$\n",
      "- Older than 60 years old: $0%$\n",
      "{'df':                  prob\n",
      "nicotine vaping   0.0\n",
      "cannabis vaping   0.0\n",
      "fashion           0.0\n",
      "entertainment     1.0\n",
      "lifestyle         0.0\n",
      "age_0-11          0.0\n",
      "age_12-17         0.0\n",
      "age_18-21         0.2\n",
      "age_22-25         0.5\n",
      "age_26-30         0.3\n",
      "age_31-40         0.0\n",
      "age_41-50         0.0\n",
      "age_51-60         0.0\n",
      "age_61_plus       0.0, 'extracted_probs': 14}\n",
      "\n",
      "Analyzing arabella_tech_9\n",
      "1406 frames read.\n",
      "Nicotine vaping: $5%$\n",
      "Cannabis vaping: $5%$\n",
      "Fashion (modeling): $5%$\n",
      "Entertainment (dj, videogames): $100%$\n",
      "Active lifestyle (sport): $5%$\n",
      "\n",
      "Age group probabilities:\n",
      "- 0-11 years old: $0%$\n",
      "- 12-17 years old: $5%$\n",
      "- 18-21 years old: $30%$\n",
      "- 22-25 years old: $40%$\n",
      "- 26-30 years old: $20%$\n",
      "- 31-40 years old: $5%$\n",
      "- 41-50 years old: $0%$\n",
      "- 51-60 years old: $0%$\n",
      "- Older than 60 years old: $0%$\n",
      "{'df':                  prob\n",
      "nicotine vaping  0.05\n",
      "cannabis vaping  0.05\n",
      "fashion          0.05\n",
      "entertainment    1.00\n",
      "lifestyle        0.05\n",
      "age_0-11         0.00\n",
      "age_12-17        0.05\n",
      "age_18-21        0.30\n",
      "age_22-25        0.40\n",
      "age_26-30        0.20\n",
      "age_31-40        0.05\n",
      "age_41-50        0.00\n",
      "age_51-60        0.00\n",
      "age_61_plus      0.00, 'extracted_probs': 14}\n",
      "\n",
      "Analyzing arabella_tech_8\n",
      "1275 frames read.\n",
      "Nicotine vaping: $0%$\n",
      "Cannabis vaping: $0%$\n",
      "Fashion (modeling): $10%$\n",
      "Entertainment (DJ, videogames): $100%$\n",
      "Active lifestyle (sport): $0%$\n",
      "\n",
      "Age 0-11: $0%$\n",
      "Age 12-17: $0%$\n",
      "Age 18-21: $20%$\n",
      "Age 22-25: $60%$\n",
      "Age 26-30: $20%$\n",
      "Age 31-40: $0%$\n",
      "Age 41-50: $0%$\n",
      "Age 51-60: $0%$\n",
      "Older than 60: $0%$\n",
      "{'df':                  prob\n",
      "nicotine vaping   0.0\n",
      "cannabis vaping   0.0\n",
      "fashion           0.1\n",
      "entertainment     1.0\n",
      "lifestyle         0.0\n",
      "age_0-11          0.0\n",
      "age_12-17         0.0\n",
      "age_18-21         0.2\n",
      "age_22-25         0.6\n",
      "age_26-30         0.2\n",
      "age_31-40         0.0\n",
      "age_41-50         0.0\n",
      "age_51-60         0.0\n",
      "age_61_plus       0.0, 'extracted_probs': 14}\n",
      "\n",
      "Analyzing arabella_tech_3\n",
      "1876 frames read.\n",
      "Nicotine vaping: $0%$\n",
      "Cannabis vaping: $0%$\n",
      "Fashion (modeling): $0%$\n",
      "Entertainment (DJ, videogames): $100%$\n",
      "Active lifestyle (sport): $0%$\n",
      "\n",
      "Age group probabilities:\n",
      "- 0-11 years old: $0%$\n",
      "- 12-17 years old: $0%$\n",
      "- 18-21 years old: $20%$\n",
      "- 22-25 years old: $60%$\n",
      "- 26-30 years old: $20%$\n",
      "- 31-40 years old: $0%$\n",
      "- 41-50 years old: $0%$\n",
      "- 51-60 years old: $0%$\n",
      "- Older than 60 years old: $0%$\n",
      "{'df':                  prob\n",
      "nicotine vaping   0.0\n",
      "cannabis vaping   0.0\n",
      "fashion           0.0\n",
      "entertainment     1.0\n",
      "lifestyle         0.0\n",
      "age_0-11          0.0\n",
      "age_12-17         0.0\n",
      "age_18-21         0.2\n",
      "age_22-25         0.6\n",
      "age_26-30         0.2\n",
      "age_31-40         0.0\n",
      "age_41-50         0.0\n",
      "age_51-60         0.0\n",
      "age_61_plus       0.0, 'extracted_probs': 14}\n",
      "\n",
      "Analyzing calitrickzz_mj_16\n",
      "657 frames read.\n",
      "- Probability of nicotine vaping: $80%$\n",
      "- Probability of cannabis vaping: $20%$\n",
      "- Probability of fashion (modeling): $10%$\n",
      "- Probability of entertainment: $30%$\n",
      "- Probability of active lifestyle (sport): $10%$\n",
      "\n",
      "- Probability of age 0-11: $0%$\n",
      "- Probability of age 12-17: $0%$\n",
      "- Probability of age 18-21: $10%$\n",
      "- Probability of age 22-25: $30%$\n",
      "- Probability of age 26-30: $40%$\n",
      "- Probability of age 31-40: $20%$\n",
      "- Probability of age 41-50: $0%$\n",
      "- Probability of age 51-60: $0%$\n",
      "- Probability of older than 60 years old: $0%$\n",
      "{'df':                  prob\n",
      "nicotine vaping   0.8\n",
      "cannabis vaping   0.2\n",
      "fashion           0.1\n",
      "entertainment     0.3\n",
      "lifestyle         0.1\n",
      "age_0-11          0.0\n",
      "age_12-17         0.0\n",
      "age_18-21         0.1\n",
      "age_22-25         0.3\n",
      "age_26-30         0.4\n",
      "age_31-40         0.2\n",
      "age_41-50         0.0\n",
      "age_51-60         0.0\n",
      "age_61_plus       0.0, 'extracted_probs': 14}\n",
      "\n",
      "Analyzing calitrickzz_mj_11\n",
      "186 frames read.\n",
      "Nicotine vaping: $0%$\n",
      "Cannabis vaping: $100%$\n",
      "Fashion (modeling): $0%$\n",
      "Entertainment: $0%$\n",
      "Active lifestyle (sport): $0%$\n",
      "\n",
      "Age group probabilities:\n",
      "0-11 years old: $0%$\n",
      "12-17 years old: $0%$\n",
      "18-21 years old: $0%$\n",
      "22-25 years old: $0%$\n",
      "26-30 years old: $0%$\n",
      "31-40 years old: $0%$\n",
      "41-50 years old: $0%$\n",
      "51-60 years old: $0%$\n",
      "Older than 60 years old: $0%$\n",
      "{'df':                  prob\n",
      "nicotine vaping   0.0\n",
      "cannabis vaping   1.0\n",
      "fashion           0.0\n",
      "entertainment     0.0\n",
      "lifestyle         0.0\n",
      "age_0-11          0.0\n",
      "age_12-17         0.0\n",
      "age_18-21         0.0\n",
      "age_22-25         0.0\n",
      "age_26-30         0.0\n",
      "age_31-40         0.0\n",
      "age_41-50         0.0\n",
      "age_51-60         0.0\n",
      "age_61_plus       0.0, 'extracted_probs': 14}\n",
      "\n",
      "Analyzing calitrickzz_mj_5\n",
      "326 frames read.\n",
      "- Probability of nicotine vaping: $90%$\n",
      "- Probability of cannabis vaping: $50%$\n",
      "- Probability of fashion (modeling): $0%$\n",
      "- Probability of entertainment: $70%$\n",
      "- Probability of active lifestyle (sport): $0%$\n",
      "- Probability of age 0-11: $0%$\n",
      "- Probability of age 12-17: $10%$\n",
      "- Probability of age 18-21: $30%$\n",
      "- Probability of age 22-25: $40%$\n",
      "- Probability of age 26-30: $15%$\n",
      "- Probability of age 31-40: $5%$\n",
      "- Probability of age 41-50: $0%$\n",
      "- Probability of age 51-60: $0%$\n",
      "- Probability of older than 60 years old: $0%$\n",
      "{'df':                  prob\n",
      "nicotine vaping  0.90\n",
      "cannabis vaping  0.50\n",
      "fashion          0.00\n",
      "entertainment    0.70\n",
      "lifestyle        0.00\n",
      "age_0-11         0.00\n",
      "age_12-17        0.10\n",
      "age_18-21        0.30\n",
      "age_22-25        0.40\n",
      "age_26-30        0.15\n",
      "age_31-40        0.05\n",
      "age_41-50        0.00\n",
      "age_51-60        0.00\n",
      "age_61_plus      0.00, 'extracted_probs': 14}\n",
      "\n",
      "Analyzing calitrickzz_mj_3_and_golf\n",
      "869 frames read.\n",
      "Nicotine vaping: $80%$\n",
      "Cannabis vaping: $70%$\n",
      "Fashion (modeling): $0%$\n",
      "Entertainment: $20%$\n",
      "Active lifestyle (sport): $90%$\n",
      "\n",
      "Age group probabilities:\n",
      "- 0-11 years old: $0%$\n",
      "- 12-17 years old: $0%$\n",
      "- 18-21 years old: $10%$\n",
      "- 22-25 years old: $20%$\n",
      "- 26-30 years old: $30%$\n",
      "- 31-40 years old: $30%$\n",
      "- 41-50 years old: $10%$\n",
      "- 51-60 years old: $0%$\n",
      "- Older than 60 years old: $0%$\n",
      "{'df':                  prob\n",
      "nicotine vaping   0.8\n",
      "cannabis vaping   0.7\n",
      "fashion           0.0\n",
      "entertainment     0.2\n",
      "lifestyle         0.9\n",
      "age_0-11          0.0\n",
      "age_12-17         0.0\n",
      "age_18-21         0.1\n",
      "age_22-25         0.2\n",
      "age_26-30         0.3\n",
      "age_31-40         0.3\n",
      "age_41-50         0.1\n",
      "age_51-60         0.0\n",
      "age_61_plus       0.0, 'extracted_probs': 14}\n",
      "\n",
      "Analyzing calitrickzz_mj_18\n",
      "813 frames read.\n",
      "Nicotine vaping: $10%$\n",
      "Cannabis vaping: $90%$\n",
      "Fashion (modeling): $0%$\n",
      "Entertainment: $20%$\n",
      "Active lifestyle (sport): $0%$\n",
      "\n",
      "Age 0-11: $0%$\n",
      "Age 12-17: $0%$\n",
      "Age 18-21: $10%$\n",
      "Age 22-25: $30%$\n",
      "Age 26-30: $40%$\n",
      "Age 31-40: $20%$\n",
      "Age 41-50: $0%$\n",
      "Age 51-60: $0%$\n",
      "Older than 60: $0%$\n",
      "{'df':                  prob\n",
      "nicotine vaping   0.1\n",
      "cannabis vaping   0.9\n",
      "fashion           0.0\n",
      "entertainment     0.2\n",
      "lifestyle         0.0\n",
      "age_0-11          0.0\n",
      "age_12-17         0.0\n",
      "age_18-21         0.1\n",
      "age_22-25         0.3\n",
      "age_26-30         0.4\n",
      "age_31-40         0.2\n",
      "age_41-50         0.0\n",
      "age_51-60         0.0\n",
      "age_61_plus       0.0, 'extracted_probs': 14}\n",
      "\n",
      "Analyzing calitrickzz_mj_13\n",
      "326 frames read.\n",
      "- Probability of nicotine vaping: $90%$\n",
      "- Probability of cannabis vaping: $10%$\n",
      "- Probability of fashion (modeling): $0%$\n",
      "- Probability of entertainment: $20%$\n",
      "- Probability of active lifestyle (sport): $0%$\n",
      "\n",
      "- Probability of age 0-11: $0%$\n",
      "- Probability of age 12-17: $10%$\n",
      "- Probability of age 18-21: $20%$\n",
      "- Probability of age 22-25: $30%$\n",
      "- Probability of age 26-30: $30%$\n",
      "- Probability of age 31-40: $10%$\n",
      "- Probability of age 41-50: $0%$\n",
      "- Probability of age 51-60: $0%$\n",
      "- Probability of older than 60 years old: $0%$\n",
      "{'df':                  prob\n",
      "nicotine vaping   0.9\n",
      "cannabis vaping   0.1\n",
      "fashion           0.0\n",
      "entertainment     0.2\n",
      "lifestyle         0.0\n",
      "age_0-11          0.0\n",
      "age_12-17         0.1\n",
      "age_18-21         0.2\n",
      "age_22-25         0.3\n",
      "age_26-30         0.3\n",
      "age_31-40         0.1\n",
      "age_41-50         0.0\n",
      "age_51-60         0.0\n",
      "age_61_plus       0.0, 'extracted_probs': 14}\n",
      "\n",
      "Analyzing calitrickzz_mj_1\n",
      "1065 frames read.\n",
      "Nicotine vaping: $0%$\n",
      "Cannabis vaping: $100%$\n",
      "Fashion (modeling): $0%$\n",
      "Entertainment: $0%$\n",
      "Active lifestyle (sport): $0%$\n",
      "\n",
      "Age group probabilities:\n",
      "0-11 years old: $0%$\n",
      "12-17 years old: $0%$\n",
      "18-21 years old: $10%$\n",
      "22-25 years old: $20%$\n",
      "26-30 years old: $30%$\n",
      "31-40 years old: $30%$\n",
      "41-50 years old: $10%$\n",
      "51-60 years old: $0%$\n",
      "Older than 60 years old: $0%$\n",
      "{'df':                  prob\n",
      "nicotine vaping   0.0\n",
      "cannabis vaping   1.0\n",
      "fashion           0.0\n",
      "entertainment     0.0\n",
      "lifestyle         0.0\n",
      "age_0-11          0.0\n",
      "age_12-17         0.0\n",
      "age_18-21         0.1\n",
      "age_22-25         0.2\n",
      "age_26-30         0.3\n",
      "age_31-40         0.3\n",
      "age_41-50         0.1\n",
      "age_51-60         0.0\n",
      "age_61_plus       0.0, 'extracted_probs': 14}\n",
      "\n",
      "Analyzing calitrickzz_mj_8\n",
      "806 frames read.\n",
      "- Probability of nicotine vaping: $10%$\n",
      "- Probability of cannabis vaping: $100%$\n",
      "- Probability of fashion (modeling): $0%$\n",
      "- Probability of entertainment (e.g., DJ, video games): $0%$\n",
      "- Probability of active lifestyle (e.g., sport): $0%$\n",
      "- Probability of age 0-11: $0%$\n",
      "- Probability of age 12-17: $0%$\n",
      "- Probability of age 18-21: $0%$\n",
      "- Probability of age 22-25: $20%$\n",
      "- Probability of age 26-30: $50%$\n",
      "- Probability of age 31-40: $30%$\n",
      "- Probability of age 41-50: $0%$\n",
      "- Probability of age 51-60: $0%$\n",
      "- Probability of older than 60 years old: $0%$\n",
      "{'df':                  prob\n",
      "nicotine vaping   0.1\n",
      "cannabis vaping   1.0\n",
      "fashion           0.0\n",
      "entertainment     0.0\n",
      "lifestyle         0.0\n",
      "age_0-11          0.0\n",
      "age_12-17         0.0\n",
      "age_18-21         0.0\n",
      "age_22-25         0.2\n",
      "age_26-30         0.5\n",
      "age_31-40         0.3\n",
      "age_41-50         0.0\n",
      "age_51-60         0.0\n",
      "age_61_plus       0.0, 'extracted_probs': 14}\n",
      "\n",
      "Analyzing calitrickzz_mj_9\n",
      "505 frames read.\n",
      "Nicotine vaping: $0%$\n",
      "Cannabis vaping: $100%$\n",
      "Fashion (modeling): $0%$\n",
      "Entertainment: $0%$\n",
      "Active lifestyle (sport): $0%$\n",
      "\n",
      "Age group probabilities:\n",
      "0-11 years old: $0%$\n",
      "12-17 years old: $0%$\n",
      "18-21 years old: $0%$\n",
      "22-25 years old: $0%$\n",
      "26-30 years old: $20%$\n",
      "31-40 years old: $70%$\n",
      "41-50 years old: $10%$\n",
      "51-60 years old: $0%$\n",
      "Older than 60 years old: $0%$\n",
      "{'df':                  prob\n",
      "nicotine vaping   0.0\n",
      "cannabis vaping   1.0\n",
      "fashion           0.0\n",
      "entertainment     0.0\n",
      "lifestyle         0.0\n",
      "age_0-11          0.0\n",
      "age_12-17         0.0\n",
      "age_18-21         0.0\n",
      "age_22-25         0.0\n",
      "age_26-30         0.2\n",
      "age_31-40         0.7\n",
      "age_41-50         0.1\n",
      "age_51-60         0.0\n",
      "age_61_plus       0.0, 'extracted_probs': 14}\n",
      "\n",
      "Analyzing calitrickzz_mj_17\n",
      "1086 frames read.\n",
      "- Nicotine vaping: $10%$\n",
      "- Cannabis vaping: $90%$\n",
      "- Fashion (modeling): $0%$\n",
      "- Entertainment: $0%$\n",
      "- Active lifestyle (sport): $0%$\n",
      "\n",
      "- Age 0-11: $0%$\n",
      "- Age 12-17: $0%$\n",
      "- Age 18-21: $10%$\n",
      "- Age 22-25: $30%$\n",
      "- Age 26-30: $40%$\n",
      "- Age 31-40: $20%$\n",
      "- Age 41-50: $0%$\n",
      "- Age 51-60: $0%$\n",
      "- Older than 60: $0%$\n",
      "{'df':                  prob\n",
      "nicotine vaping   0.1\n",
      "cannabis vaping   0.9\n",
      "fashion           0.0\n",
      "entertainment     0.0\n",
      "lifestyle         0.0\n",
      "age_0-11          0.0\n",
      "age_12-17         0.0\n",
      "age_18-21         0.1\n",
      "age_22-25         0.3\n",
      "age_26-30         0.4\n",
      "age_31-40         0.2\n",
      "age_41-50         0.0\n",
      "age_51-60         0.0\n",
      "age_61_plus       0.0, 'extracted_probs': 14}\n",
      "\n",
      "Analyzing calitrickzz_mj_10\n",
      "841 frames read.\n",
      "- Probability of nicotine vaping: $0%$\n",
      "- Probability of cannabis vaping: $100%$\n",
      "- Probability of fashion (modeling): $0%$\n",
      "- Probability of entertainment: $0%$\n",
      "- Probability of active lifestyle (sport): $0%$\n",
      "\n",
      "- Probability of age 0-11: $0%$\n",
      "- Probability of age 12-17: $0%$\n",
      "- Probability of age 18-21: $0%$\n",
      "- Probability of age 22-25: $20%$\n",
      "- Probability of age 26-30: $50%$\n",
      "- Probability of age 31-40: $30%$\n",
      "- Probability of age 41-50: $0%$\n",
      "- Probability of age 51-60: $0%$\n",
      "- Probability of older than 60 years old: $0%$\n",
      "{'df':                  prob\n",
      "nicotine vaping   0.0\n",
      "cannabis vaping   1.0\n",
      "fashion           0.0\n",
      "entertainment     0.0\n",
      "lifestyle         0.0\n",
      "age_0-11          0.0\n",
      "age_12-17         0.0\n",
      "age_18-21         0.0\n",
      "age_22-25         0.2\n",
      "age_26-30         0.5\n",
      "age_31-40         0.3\n",
      "age_41-50         0.0\n",
      "age_51-60         0.0\n",
      "age_61_plus       0.0, 'extracted_probs': 14}\n",
      "\n",
      "Analyzing calitrickzz_mj_2\n",
      "790 frames read.\n",
      "Nicotine vaping: $0%$\n",
      "Cannabis vaping: $100%$\n",
      "Fashion (modeling): $0%$\n",
      "Entertainment: $10%$\n",
      "Active lifestyle (sport): $0%$\n",
      "\n",
      "Age group probability:\n",
      "0-11 years old: $0%$\n",
      "12-17 years old: $0%$\n",
      "18-21 years old: $10%$\n",
      "22-25 years old: $20%$\n",
      "26-30 years old: $30%$\n",
      "31-40 years old: $30%$\n",
      "41-50 years old: $10%$\n",
      "51-60 years old: $0%$\n",
      "Older than 60 years old: $0%$\n",
      "{'df':                  prob\n",
      "nicotine vaping   0.0\n",
      "cannabis vaping   1.0\n",
      "fashion           0.0\n",
      "entertainment     0.1\n",
      "lifestyle         0.0\n",
      "age_0-11          0.0\n",
      "age_12-17         0.0\n",
      "age_18-21         0.1\n",
      "age_22-25         0.2\n",
      "age_26-30         0.3\n",
      "age_31-40         0.3\n",
      "age_41-50         0.1\n",
      "age_51-60         0.0\n",
      "age_61_plus       0.0, 'extracted_probs': 14}\n",
      "\n",
      "Analyzing calitrickzz_mj_20\n",
      "472 frames read.\n",
      "- Probability of nicotine vaping: $0%$\n",
      "- Probability of cannabis vaping: $100%$\n",
      "- Probability of fashion (modeling): $0%$\n",
      "- Probability of entertainment: $0%$\n",
      "- Probability of active lifestyle (sport): $0%$\n",
      "- Probability of age 0-11: $0%$\n",
      "- Probability of age 12-17: $0%$\n",
      "- Probability of age 18-21: $0%$\n",
      "- Probability of age 22-25: $50%$\n",
      "- Probability of age 26-30: $50%$\n",
      "- Probability of age 31-40: $0%$\n",
      "- Probability of age 41-50: $0%$\n",
      "- Probability of age 51-60: $0%$\n",
      "- Probability of older than 60 years old: $0%$\n",
      "{'df':                  prob\n",
      "nicotine vaping   0.0\n",
      "cannabis vaping   1.0\n",
      "fashion           0.0\n",
      "entertainment     0.0\n",
      "lifestyle         0.0\n",
      "age_0-11          0.0\n",
      "age_12-17         0.0\n",
      "age_18-21         0.0\n",
      "age_22-25         0.5\n",
      "age_26-30         0.5\n",
      "age_31-40         0.0\n",
      "age_41-50         0.0\n",
      "age_51-60         0.0\n",
      "age_61_plus       0.0, 'extracted_probs': 14}\n",
      "\n",
      "Analyzing calitrickzz_mj_12\n",
      "490 frames read.\n",
      "- Probability of nicotine vaping: $10%$\n",
      "- Probability of cannabis vaping: $100%$\n",
      "- Probability of fashion (modeling): $0%$\n",
      "- Probability of entertainment: $0%$\n",
      "- Probability of active lifestyle (sport): $0%$\n",
      "- Probability of age 0-11: $0%$\n",
      "- Probability of age 12-17: $0%$\n",
      "- Probability of age 18-21: $10%$\n",
      "- Probability of age 22-25: $30%$\n",
      "- Probability of age 26-30: $40%$\n",
      "- Probability of age 31-40: $20%$\n",
      "- Probability of age 41-50: $0%$\n",
      "- Probability of age 51-60: $0%$\n",
      "- Probability of older than 60 years old: $0%$\n",
      "{'df':                  prob\n",
      "nicotine vaping   0.1\n",
      "cannabis vaping   1.0\n",
      "fashion           0.0\n",
      "entertainment     0.0\n",
      "lifestyle         0.0\n",
      "age_0-11          0.0\n",
      "age_12-17         0.0\n",
      "age_18-21         0.1\n",
      "age_22-25         0.3\n",
      "age_26-30         0.4\n",
      "age_31-40         0.2\n",
      "age_41-50         0.0\n",
      "age_51-60         0.0\n",
      "age_61_plus       0.0, 'extracted_probs': 14}\n",
      "\n",
      "Analyzing calitrickzz_mj_6\n",
      "798 frames read.\n",
      "Nicotine vaping: $10%$\n",
      "Cannabis vaping: $100%$\n",
      "Fashion (modeling): $0%$\n",
      "Entertainment: $0%$\n",
      "Active lifestyle (sport): $0%$\n",
      "\n",
      "Age group probabilities:\n",
      "0-11 years old: $0%$\n",
      "12-17 years old: $10%$\n",
      "18-21 years old: $20%$\n",
      "22-25 years old: $30%$\n",
      "26-30 years old: $20%$\n",
      "31-40 years old: $10%$\n",
      "41-50 years old: $5%$\n",
      "51-60 years old: $5%$\n",
      "Older than 60 years old: $0%$\n",
      "{'df':                  prob\n",
      "nicotine vaping  0.10\n",
      "cannabis vaping  1.00\n",
      "fashion          0.00\n",
      "entertainment    0.00\n",
      "lifestyle        0.00\n",
      "age_0-11         0.00\n",
      "age_12-17        0.10\n",
      "age_18-21        0.20\n",
      "age_22-25        0.30\n",
      "age_26-30        0.20\n",
      "age_31-40        0.10\n",
      "age_41-50        0.05\n",
      "age_51-60        0.05\n",
      "age_61_plus      0.00, 'extracted_probs': 14}\n",
      "\n",
      "Analyzing calitrickzz_mj_14\n",
      "342 frames read.\n",
      "Nicotine vaping: $90%$\n",
      "Cannabis vaping: $80%$\n",
      "Fashion (modeling): $0%$\n",
      "Entertainment: $20%$\n",
      "Active lifestyle (sport): $0%$\n",
      "\n",
      "Age 0-11: $0%$\n",
      "Age 12-17: $10%$\n",
      "Age 18-21: $20%$\n",
      "Age 22-25: $30%$\n",
      "Age 26-30: $30%$\n",
      "Age 31-40: $10%$\n",
      "Age 41-50: $0%$\n",
      "Age 51-60: $0%$\n",
      "Older than 60: $0%$\n",
      "{'df':                  prob\n",
      "nicotine vaping   0.9\n",
      "cannabis vaping   0.8\n",
      "fashion           0.0\n",
      "entertainment     0.2\n",
      "lifestyle         0.0\n",
      "age_0-11          0.0\n",
      "age_12-17         0.1\n",
      "age_18-21         0.2\n",
      "age_22-25         0.3\n",
      "age_26-30         0.3\n",
      "age_31-40         0.1\n",
      "age_41-50         0.0\n",
      "age_51-60         0.0\n",
      "age_61_plus       0.0, 'extracted_probs': 14}\n",
      "\n",
      "Analyzing calitrickzz_mj_19\n",
      "231 frames read.\n",
      "Nicotine vaping: $80%$\n",
      "Cannabis vaping: $80%$\n",
      "Fashion (modeling): $10%$\n",
      "Entertainment: $60%$\n",
      "Active lifestyle (sport): $10%$\n",
      "\n",
      "Age 0-11: $0%$\n",
      "Age 12-17: $0%$\n",
      "Age 18-21: $10%$\n",
      "Age 22-25: $30%$\n",
      "Age 26-30: $40%$\n",
      "Age 31-40: $20%$\n",
      "Age 41-50: $0%$\n",
      "Age 51-60: $0%$\n",
      "Older than 60: $0%$\n",
      "{'df':                  prob\n",
      "nicotine vaping   0.8\n",
      "cannabis vaping   0.8\n",
      "fashion           0.1\n",
      "entertainment     0.6\n",
      "lifestyle         0.1\n",
      "age_0-11          0.0\n",
      "age_12-17         0.0\n",
      "age_18-21         0.1\n",
      "age_22-25         0.3\n",
      "age_26-30         0.4\n",
      "age_31-40         0.2\n",
      "age_41-50         0.0\n",
      "age_51-60         0.0\n",
      "age_61_plus       0.0, 'extracted_probs': 14}\n",
      "\n",
      "Analyzing calitrickzz_mj_4\n",
      "703 frames read.\n",
      "- Probability of nicotine vaping: $80%$\n",
      "- Probability of cannabis vaping: $70%$\n",
      "- Probability of fashion (modeling): $10%$\n",
      "- Probability of entertainment: $20%$\n",
      "- Probability of active lifestyle (sport): $10%$\n",
      "- Probability of age 0-11: $0%$\n",
      "- Probability of age 12-17: $0%$\n",
      "- Probability of age 18-21: $20%$\n",
      "- Probability of age 22-25: $40%$\n",
      "- Probability of age 26-30: $30%$\n",
      "- Probability of age 31-40: $10%$\n",
      "- Probability of age 41-50: $0%$\n",
      "- Probability of age 51-60: $0%$\n",
      "- Probability of older than 60 years old: $0%$\n",
      "{'df':                  prob\n",
      "nicotine vaping   0.8\n",
      "cannabis vaping   0.7\n",
      "fashion           0.1\n",
      "entertainment     0.2\n",
      "lifestyle         0.1\n",
      "age_0-11          0.0\n",
      "age_12-17         0.0\n",
      "age_18-21         0.2\n",
      "age_22-25         0.4\n",
      "age_26-30         0.3\n",
      "age_31-40         0.1\n",
      "age_41-50         0.0\n",
      "age_51-60         0.0\n",
      "age_61_plus       0.0, 'extracted_probs': 14}\n",
      "\n",
      "Analyzing calitrickzz_mj_7\n",
      "943 frames read.\n",
      "Nicotine vaping: $0%$\n",
      "Cannabis vaping: $100%$\n",
      "Fashion (modeling): $0%$\n",
      "Entertainment: $0%$\n",
      "Active lifestyle (sport): $0%$\n",
      "\n",
      "Age 0-11: $0%$\n",
      "Age 12-17: $0%$\n",
      "Age 18-21: $0%$\n",
      "Age 22-25: $0%$\n",
      "Age 26-30: $0%$\n",
      "Age 31-40: $0%$\n",
      "Age 41-50: $0%$\n",
      "Age 51-60: $0%$\n",
      "Older than 60: $0%$\n",
      "{'df':                  prob\n",
      "nicotine vaping   0.0\n",
      "cannabis vaping   1.0\n",
      "fashion           0.0\n",
      "entertainment     0.0\n",
      "lifestyle         0.0\n",
      "age_0-11          0.0\n",
      "age_12-17         0.0\n",
      "age_18-21         0.0\n",
      "age_22-25         0.0\n",
      "age_26-30         0.0\n",
      "age_31-40         0.0\n",
      "age_41-50         0.0\n",
      "age_51-60         0.0\n",
      "age_61_plus       0.0, 'extracted_probs': 14}\n",
      "\n",
      "Analyzing calitrickzz_mj_15\n",
      "835 frames read.\n",
      "- Probability of nicotine vaping: $10%$\n",
      "- Probability of cannabis vaping: $90%$\n",
      "- Probability of fashion (modeling): $0%$\n",
      "- Probability of entertainment (dj, videogames): $0%$\n",
      "- Probability of active lifestyle (sport): $0%$\n",
      "- Probability of age 0-11: $0%$\n",
      "- Probability of age 12-17: $0%$\n",
      "- Probability of age 18-21: $20%$\n",
      "- Probability of age 22-25: $50%$\n",
      "- Probability of age 26-30: $20%$\n",
      "- Probability of age 31-40: $10%$\n",
      "- Probability of age 41-50: $0%$\n",
      "- Probability of age 51-60: $0%$\n",
      "- Probability of older than 60 years old: $0%$\n",
      "{'df':                  prob\n",
      "nicotine vaping   0.1\n",
      "cannabis vaping   0.9\n",
      "fashion           0.0\n",
      "entertainment     0.0\n",
      "lifestyle         0.0\n",
      "age_0-11          0.0\n",
      "age_12-17         0.0\n",
      "age_18-21         0.2\n",
      "age_22-25         0.5\n",
      "age_26-30         0.2\n",
      "age_31-40         0.1\n",
      "age_41-50         0.0\n",
      "age_51-60         0.0\n",
      "age_61_plus       0.0, 'extracted_probs': 14}\n",
      "CPU times: user 38min 39s, sys: 5.42 s, total: 38min 44s\n",
      "Wall time: 47min 16s\n"
     ]
    }
   ],
   "source": [
    "%%time\n",
    "\n",
    "results_probs = analyze_videos_probs(videos, prompt_probs, verbose = True)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 335,
   "id": "a3068f59-3ead-4ceb-ab9e-c858c515b7b9",
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "102"
      ]
     },
     "execution_count": 335,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "len(results_probs)"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "4a6b1c55-e063-4d38-89e4-8b2ae7a9c92f",
   "metadata": {},
   "source": [
    "## Convert extracted probabilities into a complete dataframe"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 336,
   "id": "c5f01c27-d690-4536-93a2-d9685bf0fbed",
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "dict_keys(['fabio_fashion_3', 'chamillioneyes_fashion_1', 'fabio_fashion_2', 'fabio_fashion_18', 'fabio_fashion_19', 'fabio_fashion_14', 'fabio_fashion_4', 'fabio_fashion_17', 'fabio_fashion_8', 'fabio_fashion_9', 'fabio_fashion_12', 'fabio_fashion_1', 'fabio_fashion_16', 'fabio_fashion_11', 'fabio_fashion_7', 'fabio_fashion_15', 'fabio_fashion_6', 'fabio_fashion_13', 'fabio_fashion_10_andpod', 'fabio_fashion_5', 'fabio_health_3', 'drewdirps_health_2', 'fabio_health_2_and_pod', 'arabella_health_1', 'chamil_health_vape', 'chamillioneyes_health_1', 'edripps_health_vape', 'drewdirps_health_4', 'drewdirps_health_1', 'vapes_aby_health_1_andvape', 'vapes_aby_health_2_andvape', 'alex_health_vape', 'drew_health_vape', 'chamillioneyes_health_2', 'vapes_aby_health_vape', 'fabio_health_1', 'drewdirps_health_3', 'alex_health_2', 'arabella_health_2', 'alex_health_1', 'drewdirps_ecigs_3', 'chamillioneyes_ecgis_1', 'drewdirps_ecigs_4', 'edripss_ecgis_2', 'chamillioneyes_ecigs_2', 'calitrickzz_ecigs_3', 'vapes_aby_ecigs_1', 'edripss_ecgis_1', 'vapes_aby_ecigs_2', 'arabella_ecigs_4', 'arabella_ecigs_2', 'drewdirps_ecigs_1', 'arabella_ecigs_5', 'calitrickzz_ecigs_2', 'alex_ecigs_1', 'fabio_ecigs_1', 'drewdirps_ecigs_2', 'arabella_ecigs_1', 'zanxlyfe_ecigs_1', 'arabella_ecigs_3', 'calitrickzz_ecigs_1', 'zanxlyfe_ecigs_2', 'zanxlyfe_tech_1', 'arabella_tech_2', 'zanxlyfe_tech_5', 'zanxlyfe_tech_6', 'zanxlyfe_tech_10', 'zanxlyfe_tech_4', 'arabella_tech_4', 'zanxlyfe_tech_8', 'arabella_tech_7', 'arabella_tech_10', 'arabella_tech_1', 'zanxlyfe_tech_3', 'zanxlyfe_tech_7', 'zanxlyfe_tech_2', 'arabella_tech_5', 'arabella_tech_6', 'zanxlyfe_tech_9', 'arabella_tech_9', 'arabella_tech_8', 'arabella_tech_3', 'calitrickzz_mj_16', 'calitrickzz_mj_11', 'calitrickzz_mj_5', 'calitrickzz_mj_3_and_golf', 'calitrickzz_mj_18', 'calitrickzz_mj_13', 'calitrickzz_mj_1', 'calitrickzz_mj_8', 'calitrickzz_mj_9', 'calitrickzz_mj_17', 'calitrickzz_mj_10', 'calitrickzz_mj_2', 'calitrickzz_mj_20', 'calitrickzz_mj_12', 'calitrickzz_mj_6', 'calitrickzz_mj_14', 'calitrickzz_mj_19', 'calitrickzz_mj_4', 'calitrickzz_mj_7', 'calitrickzz_mj_15'])"
      ]
     },
     "execution_count": 336,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "results_probs.keys()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 337,
   "id": "e7304278-f932-40c3-99a1-01244edc6f07",
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "{'query': ChatCompletion(id='chatcmpl-9R1B9ekPvqLALnLhiPPPC1jLR3kig', choices=[Choice(finish_reason='stop', index=0, logprobs=None, message=ChatCompletionMessage(content='Nicotine vaping: $5%$\\nCannabis vaping: $5%$\\nFashion (modeling): $80%$\\nEntertainment: $10%$\\nActive lifestyle (sport): $10%$\\n\\nAge group probabilities:\\n0-11 years old: $0%$\\n12-17 years old: $5%$\\n18-21 years old: $20%$\\n22-25 years old: $40%$\\n26-30 years old: $25%$\\n31-40 years old: $10%$\\n41-50 years old: $0%$\\n51-60 years old: $0%$\\nOlder than 60 years old: $0%$', role='assistant', function_call=None, tool_calls=None))], created=1716226415, model='gpt-4o-2024-05-13', object='chat.completion', system_fingerprint='fp_927397958d', usage=CompletionUsage(completion_tokens=135, prompt_tokens=3924, total_tokens=4059)),\n",
       " 'msg': 'Nicotine vaping: $5%$\\nCannabis vaping: $5%$\\nFashion (modeling): $80%$\\nEntertainment: $10%$\\nActive lifestyle (sport): $10%$\\n\\nAge group probabilities:\\n0-11 years old: $0%$\\n12-17 years old: $5%$\\n18-21 years old: $20%$\\n22-25 years old: $40%$\\n26-30 years old: $25%$\\n31-40 years old: $10%$\\n41-50 years old: $0%$\\n51-60 years old: $0%$\\nOlder than 60 years old: $0%$',\n",
       " 'parsed': {'df':                  prob\n",
       "  nicotine vaping  0.05\n",
       "  cannabis vaping  0.05\n",
       "  fashion          0.80\n",
       "  entertainment    0.10\n",
       "  lifestyle        0.10\n",
       "  age_0-11         0.00\n",
       "  age_12-17        0.05\n",
       "  age_18-21        0.20\n",
       "  age_22-25        0.40\n",
       "  age_26-30        0.25\n",
       "  age_31-40        0.10\n",
       "  age_41-50        0.00\n",
       "  age_51-60        0.00\n",
       "  age_61_plus      0.00,\n",
       "  'extracted_probs': 14},\n",
       " 'tries': 1}"
      ]
     },
     "execution_count": 337,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "results_probs['fabio_fashion_3']"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 259,
   "id": "8a7cf128-994e-4474-ad55-4969756883fb",
   "metadata": {},
   "outputs": [],
   "source": [
    "video_name = 'fabio_fashion_3'\n",
    "cur_result = results_probs[video_name]\n",
    "cur_df = cur_result['parsed']['df']"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 264,
   "id": "0e9a6e35-4f0d-45a6-b87e-4648a3eb2ecc",
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/html": [
       "<div>\n",
       "<style scoped>\n",
       "    .dataframe tbody tr th:only-of-type {\n",
       "        vertical-align: middle;\n",
       "    }\n",
       "\n",
       "    .dataframe tbody tr th {\n",
       "        vertical-align: top;\n",
       "    }\n",
       "\n",
       "    .dataframe thead th {\n",
       "        text-align: right;\n",
       "    }\n",
       "</style>\n",
       "<table border=\"1\" class=\"dataframe\">\n",
       "  <thead>\n",
       "    <tr style=\"text-align: right;\">\n",
       "      <th></th>\n",
       "      <th>video</th>\n",
       "      <th>extracted_probs</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>0</th>\n",
       "      <td>fabio_fashion_3</td>\n",
       "      <td>7</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "</div>"
      ],
      "text/plain": [
       "             video  extracted_probs\n",
       "0  fabio_fashion_3                7"
      ]
     },
     "execution_count": 264,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "extra_df = pd.DataFrame({'video': video_name,\n",
    "                         'tries': cur_result['tries'],\n",
    "                         'extracted_probs': cur_result['parsed']['extracted_probs']\n",
    "                        }, index = [0])\n",
    "extra_df"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 265,
   "id": "b0c65975-311b-4705-8ede-ec61e15e5d2a",
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/html": [
       "<div>\n",
       "<style scoped>\n",
       "    .dataframe tbody tr th:only-of-type {\n",
       "        vertical-align: middle;\n",
       "    }\n",
       "\n",
       "    .dataframe tbody tr th {\n",
       "        vertical-align: top;\n",
       "    }\n",
       "\n",
       "    .dataframe thead th {\n",
       "        text-align: right;\n",
       "    }\n",
       "</style>\n",
       "<table border=\"1\" class=\"dataframe\">\n",
       "  <thead>\n",
       "    <tr style=\"text-align: right;\">\n",
       "      <th></th>\n",
       "      <th>nicotine vaping</th>\n",
       "      <th>cannabis vaping</th>\n",
       "      <th>fashion</th>\n",
       "      <th>entertainment</th>\n",
       "      <th>lifestyle</th>\n",
       "      <th>age_0-11</th>\n",
       "      <th>age_12-17</th>\n",
       "      <th>age_18-21</th>\n",
       "      <th>age_22-25</th>\n",
       "      <th>age_26-30</th>\n",
       "      <th>age_31-40</th>\n",
       "      <th>age_41-50</th>\n",
       "      <th>age_51-60</th>\n",
       "      <th>age_61_plus</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>prob</th>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>1.0</td>\n",
       "      <td>0.1</td>\n",
       "      <td>0.1</td>\n",
       "      <td>None</td>\n",
       "      <td>None</td>\n",
       "      <td>None</td>\n",
       "      <td>0.5</td>\n",
       "      <td>0.5</td>\n",
       "      <td>None</td>\n",
       "      <td>None</td>\n",
       "      <td>None</td>\n",
       "      <td>None</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "</div>"
      ],
      "text/plain": [
       "     nicotine vaping cannabis vaping fashion entertainment lifestyle age_0-11  \\\n",
       "prob             0.0             0.0     1.0           0.1       0.1     None   \n",
       "\n",
       "     age_12-17 age_18-21 age_22-25 age_26-30 age_31-40 age_41-50 age_51-60  \\\n",
       "prob      None      None       0.5       0.5      None      None      None   \n",
       "\n",
       "     age_61_plus  \n",
       "prob        None  "
      ]
     },
     "execution_count": 265,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "t_df = cur_df.T\n",
    "t_df"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 266,
   "id": "e801b8a6-ec90-40a9-9896-91becd7bf0f1",
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/html": [
       "<div>\n",
       "<style scoped>\n",
       "    .dataframe tbody tr th:only-of-type {\n",
       "        vertical-align: middle;\n",
       "    }\n",
       "\n",
       "    .dataframe tbody tr th {\n",
       "        vertical-align: top;\n",
       "    }\n",
       "\n",
       "    .dataframe thead th {\n",
       "        text-align: right;\n",
       "    }\n",
       "</style>\n",
       "<table border=\"1\" class=\"dataframe\">\n",
       "  <thead>\n",
       "    <tr style=\"text-align: right;\">\n",
       "      <th></th>\n",
       "      <th>video</th>\n",
       "      <th>extracted_probs</th>\n",
       "      <th>nicotine vaping</th>\n",
       "      <th>cannabis vaping</th>\n",
       "      <th>fashion</th>\n",
       "      <th>entertainment</th>\n",
       "      <th>lifestyle</th>\n",
       "      <th>age_0-11</th>\n",
       "      <th>age_12-17</th>\n",
       "      <th>age_18-21</th>\n",
       "      <th>age_22-25</th>\n",
       "      <th>age_26-30</th>\n",
       "      <th>age_31-40</th>\n",
       "      <th>age_41-50</th>\n",
       "      <th>age_51-60</th>\n",
       "      <th>age_61_plus</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>0</th>\n",
       "      <td>fabio_fashion_3</td>\n",
       "      <td>7</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>1.0</td>\n",
       "      <td>0.1</td>\n",
       "      <td>0.1</td>\n",
       "      <td>None</td>\n",
       "      <td>None</td>\n",
       "      <td>None</td>\n",
       "      <td>0.5</td>\n",
       "      <td>0.5</td>\n",
       "      <td>None</td>\n",
       "      <td>None</td>\n",
       "      <td>None</td>\n",
       "      <td>None</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "</div>"
      ],
      "text/plain": [
       "             video  extracted_probs nicotine vaping cannabis vaping fashion  \\\n",
       "0  fabio_fashion_3                7             0.0             0.0     1.0   \n",
       "\n",
       "  entertainment lifestyle age_0-11 age_12-17 age_18-21 age_22-25 age_26-30  \\\n",
       "0           0.1       0.1     None      None      None       0.5       0.5   \n",
       "\n",
       "  age_31-40 age_41-50 age_51-60 age_61_plus  \n",
       "0      None      None      None        None  "
      ]
     },
     "execution_count": 266,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "new_df = pd.concat([extra_df, cur_df.T.reset_index(drop = True)], axis = 1, join = 'outer')\n",
    "new_df"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 338,
   "id": "9fce1331-f917-4994-95dd-7563d1029e1a",
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "102"
      ]
     },
     "execution_count": 338,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "new_dfs = []\n",
    "for video_name, result in results_probs.items():\n",
    "    cur_df = result['parsed']['df']\n",
    "    extra_df = pd.DataFrame({'video': video_name,\n",
    "                             'tries': result['tries'],\n",
    "                             'extracted_probs': result['parsed']['extracted_probs']\n",
    "                        }, index = [0])\n",
    "    new_df = pd.concat([extra_df, cur_df.T.reset_index(drop = True)], axis = 1, join = 'outer')\n",
    "    new_dfs.append(new_df)\n",
    "len(new_dfs)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 339,
   "id": "89828887-7aa4-4b31-a9b7-2470f89feaa8",
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "(102, 17)"
      ]
     },
     "execution_count": 339,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "result_df = pd.concat(new_dfs, axis = 0)\n",
    "result_df.shape"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 340,
   "id": "c7a06f2c-dd88-4d57-9048-492f3bd4164a",
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/html": [
       "<div>\n",
       "<style scoped>\n",
       "    .dataframe tbody tr th:only-of-type {\n",
       "        vertical-align: middle;\n",
       "    }\n",
       "\n",
       "    .dataframe tbody tr th {\n",
       "        vertical-align: top;\n",
       "    }\n",
       "\n",
       "    .dataframe thead th {\n",
       "        text-align: right;\n",
       "    }\n",
       "</style>\n",
       "<table border=\"1\" class=\"dataframe\">\n",
       "  <thead>\n",
       "    <tr style=\"text-align: right;\">\n",
       "      <th></th>\n",
       "      <th>video</th>\n",
       "      <th>tries</th>\n",
       "      <th>extracted_probs</th>\n",
       "      <th>nicotine vaping</th>\n",
       "      <th>cannabis vaping</th>\n",
       "      <th>fashion</th>\n",
       "      <th>entertainment</th>\n",
       "      <th>lifestyle</th>\n",
       "      <th>age_0-11</th>\n",
       "      <th>age_12-17</th>\n",
       "      <th>age_18-21</th>\n",
       "      <th>age_22-25</th>\n",
       "      <th>age_26-30</th>\n",
       "      <th>age_31-40</th>\n",
       "      <th>age_41-50</th>\n",
       "      <th>age_51-60</th>\n",
       "      <th>age_61_plus</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>0</th>\n",
       "      <td>fabio_fashion_3</td>\n",
       "      <td>1</td>\n",
       "      <td>14</td>\n",
       "      <td>0.05</td>\n",
       "      <td>0.05</td>\n",
       "      <td>0.8</td>\n",
       "      <td>0.1</td>\n",
       "      <td>0.1</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.05</td>\n",
       "      <td>0.2</td>\n",
       "      <td>0.4</td>\n",
       "      <td>0.25</td>\n",
       "      <td>0.1</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>0</th>\n",
       "      <td>chamillioneyes_fashion_1</td>\n",
       "      <td>1</td>\n",
       "      <td>14</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.8</td>\n",
       "      <td>0.2</td>\n",
       "      <td>0.1</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.2</td>\n",
       "      <td>0.5</td>\n",
       "      <td>0.2</td>\n",
       "      <td>0.1</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>0</th>\n",
       "      <td>fabio_fashion_2</td>\n",
       "      <td>1</td>\n",
       "      <td>14</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.8</td>\n",
       "      <td>0.1</td>\n",
       "      <td>0.1</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.1</td>\n",
       "      <td>0.3</td>\n",
       "      <td>0.4</td>\n",
       "      <td>0.2</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>0</th>\n",
       "      <td>fabio_fashion_18</td>\n",
       "      <td>1</td>\n",
       "      <td>14</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.8</td>\n",
       "      <td>0.1</td>\n",
       "      <td>0.1</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.1</td>\n",
       "      <td>0.4</td>\n",
       "      <td>0.3</td>\n",
       "      <td>0.2</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>0</th>\n",
       "      <td>fabio_fashion_19</td>\n",
       "      <td>1</td>\n",
       "      <td>14</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.8</td>\n",
       "      <td>0.1</td>\n",
       "      <td>0.1</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.1</td>\n",
       "      <td>0.4</td>\n",
       "      <td>0.4</td>\n",
       "      <td>0.1</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "</div>"
      ],
      "text/plain": [
       "                      video  tries  extracted_probs nicotine vaping  \\\n",
       "0           fabio_fashion_3      1               14            0.05   \n",
       "0  chamillioneyes_fashion_1      1               14             0.0   \n",
       "0           fabio_fashion_2      1               14             0.0   \n",
       "0          fabio_fashion_18      1               14             0.0   \n",
       "0          fabio_fashion_19      1               14             0.0   \n",
       "\n",
       "  cannabis vaping fashion entertainment lifestyle age_0-11 age_12-17  \\\n",
       "0            0.05     0.8           0.1       0.1      0.0      0.05   \n",
       "0             0.0     0.8           0.2       0.1      0.0       0.0   \n",
       "0             0.0     0.8           0.1       0.1      0.0       0.0   \n",
       "0             0.0     0.8           0.1       0.1      0.0       0.0   \n",
       "0             0.0     0.8           0.1       0.1      0.0       0.0   \n",
       "\n",
       "  age_18-21 age_22-25 age_26-30 age_31-40 age_41-50 age_51-60 age_61_plus  \n",
       "0       0.2       0.4      0.25       0.1       0.0       0.0         0.0  \n",
       "0       0.2       0.5       0.2       0.1       0.0       0.0         0.0  \n",
       "0       0.1       0.3       0.4       0.2       0.0       0.0         0.0  \n",
       "0       0.1       0.4       0.3       0.2       0.0       0.0         0.0  \n",
       "0       0.1       0.4       0.4       0.1       0.0       0.0         0.0  "
      ]
     },
     "execution_count": 340,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "result_df.head()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 341,
   "id": "76fa8494-49c0-4223-bd04-24e1fde54c19",
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "video              1.000000\n",
       "tries              1.000000\n",
       "extracted_probs    1.000000\n",
       "nicotine vaping    1.000000\n",
       "cannabis vaping    1.000000\n",
       "fashion            1.000000\n",
       "entertainment      1.000000\n",
       "lifestyle          1.000000\n",
       "age_0-11           0.990196\n",
       "age_12-17          0.990196\n",
       "age_18-21          0.990196\n",
       "age_22-25          1.000000\n",
       "age_26-30          1.000000\n",
       "age_31-40          0.990196\n",
       "age_41-50          0.970588\n",
       "age_51-60          0.960784\n",
       "age_61_plus        0.960784\n",
       "dtype: float64"
      ]
     },
     "execution_count": 341,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "# Non-missing percentages for each column\n",
    "1 - result_df.isnull().mean()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 342,
   "id": "945e4c68-96cd-49ae-9060-80da30cbcdac",
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "tries\n",
       "1    102\n",
       "Name: count, dtype: int64"
      ]
     },
     "execution_count": 342,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "# Everything succeeded in one try - sweet.\n",
    "result_df.tries.value_counts()"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "be764684-e561-44ad-9cc4-cc395c83b0a8",
   "metadata": {},
   "source": [
    "### Merge in labels"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 343,
   "id": "d6e1bd4b-874a-418a-a973-8388beda3e72",
   "metadata": {},
   "outputs": [],
   "source": [
    "df['video'] = df.path.str.extract(r'/([^/]+).mp4$')"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 344,
   "id": "06d87097-e802-4652-86f8-419287d093f2",
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/html": [
       "<div>\n",
       "<style scoped>\n",
       "    .dataframe tbody tr th:only-of-type {\n",
       "        vertical-align: middle;\n",
       "    }\n",
       "\n",
       "    .dataframe tbody tr th {\n",
       "        vertical-align: top;\n",
       "    }\n",
       "\n",
       "    .dataframe thead th {\n",
       "        text-align: right;\n",
       "    }\n",
       "</style>\n",
       "<table border=\"1\" class=\"dataframe\">\n",
       "  <thead>\n",
       "    <tr style=\"text-align: right;\">\n",
       "      <th></th>\n",
       "      <th>path</th>\n",
       "      <th>theme</th>\n",
       "      <th>video</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>0</th>\n",
       "      <td>videos/GPT4_themes/fashion/fabio_fashion_3.mp4</td>\n",
       "      <td>fashion</td>\n",
       "      <td>fabio_fashion_3</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>1</th>\n",
       "      <td>videos/GPT4_themes/fashion/chamillioneyes_fash...</td>\n",
       "      <td>fashion</td>\n",
       "      <td>chamillioneyes_fashion_1</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>2</th>\n",
       "      <td>videos/GPT4_themes/fashion/fabio_fashion_2.mp4</td>\n",
       "      <td>fashion</td>\n",
       "      <td>fabio_fashion_2</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>3</th>\n",
       "      <td>videos/GPT4_themes/fashion/fabio_fashion_18.mp4</td>\n",
       "      <td>fashion</td>\n",
       "      <td>fabio_fashion_18</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>4</th>\n",
       "      <td>videos/GPT4_themes/fashion/fabio_fashion_19.mp4</td>\n",
       "      <td>fashion</td>\n",
       "      <td>fabio_fashion_19</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "</div>"
      ],
      "text/plain": [
       "                                                path    theme  \\\n",
       "0     videos/GPT4_themes/fashion/fabio_fashion_3.mp4  fashion   \n",
       "1  videos/GPT4_themes/fashion/chamillioneyes_fash...  fashion   \n",
       "2     videos/GPT4_themes/fashion/fabio_fashion_2.mp4  fashion   \n",
       "3    videos/GPT4_themes/fashion/fabio_fashion_18.mp4  fashion   \n",
       "4    videos/GPT4_themes/fashion/fabio_fashion_19.mp4  fashion   \n",
       "\n",
       "                      video  \n",
       "0           fabio_fashion_3  \n",
       "1  chamillioneyes_fashion_1  \n",
       "2           fabio_fashion_2  \n",
       "3          fabio_fashion_18  \n",
       "4          fabio_fashion_19  "
      ]
     },
     "execution_count": 344,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "df.head()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 345,
   "id": "3821e386-b53a-458c-be05-134e3e3fb825",
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "theme\n",
       "ecigs         22\n",
       "fashion       20\n",
       "health        20\n",
       "technology    20\n",
       "marijuana     20\n",
       "Name: count, dtype: int64"
      ]
     },
     "execution_count": 345,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "df.theme.value_counts()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 346,
   "id": "0eaba380-3cf9-4007-bebe-46a7b33f4278",
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/html": [
       "<div>\n",
       "<style scoped>\n",
       "    .dataframe tbody tr th:only-of-type {\n",
       "        vertical-align: middle;\n",
       "    }\n",
       "\n",
       "    .dataframe tbody tr th {\n",
       "        vertical-align: top;\n",
       "    }\n",
       "\n",
       "    .dataframe thead th {\n",
       "        text-align: right;\n",
       "    }\n",
       "</style>\n",
       "<table border=\"1\" class=\"dataframe\">\n",
       "  <thead>\n",
       "    <tr style=\"text-align: right;\">\n",
       "      <th></th>\n",
       "      <th>path</th>\n",
       "      <th>theme</th>\n",
       "      <th>video</th>\n",
       "      <th>tries</th>\n",
       "      <th>extracted_probs</th>\n",
       "      <th>nicotine vaping</th>\n",
       "      <th>cannabis vaping</th>\n",
       "      <th>fashion</th>\n",
       "      <th>entertainment</th>\n",
       "      <th>lifestyle</th>\n",
       "      <th>age_0-11</th>\n",
       "      <th>age_12-17</th>\n",
       "      <th>age_18-21</th>\n",
       "      <th>age_22-25</th>\n",
       "      <th>age_26-30</th>\n",
       "      <th>age_31-40</th>\n",
       "      <th>age_41-50</th>\n",
       "      <th>age_51-60</th>\n",
       "      <th>age_61_plus</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>0</th>\n",
       "      <td>videos/GPT4_themes/fashion/fabio_fashion_3.mp4</td>\n",
       "      <td>fashion</td>\n",
       "      <td>fabio_fashion_3</td>\n",
       "      <td>1</td>\n",
       "      <td>14</td>\n",
       "      <td>0.05</td>\n",
       "      <td>0.05</td>\n",
       "      <td>0.8</td>\n",
       "      <td>0.1</td>\n",
       "      <td>0.1</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.05</td>\n",
       "      <td>0.2</td>\n",
       "      <td>0.4</td>\n",
       "      <td>0.25</td>\n",
       "      <td>0.1</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>1</th>\n",
       "      <td>videos/GPT4_themes/fashion/chamillioneyes_fash...</td>\n",
       "      <td>fashion</td>\n",
       "      <td>chamillioneyes_fashion_1</td>\n",
       "      <td>1</td>\n",
       "      <td>14</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.8</td>\n",
       "      <td>0.2</td>\n",
       "      <td>0.1</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.2</td>\n",
       "      <td>0.5</td>\n",
       "      <td>0.2</td>\n",
       "      <td>0.1</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>2</th>\n",
       "      <td>videos/GPT4_themes/fashion/fabio_fashion_2.mp4</td>\n",
       "      <td>fashion</td>\n",
       "      <td>fabio_fashion_2</td>\n",
       "      <td>1</td>\n",
       "      <td>14</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.8</td>\n",
       "      <td>0.1</td>\n",
       "      <td>0.1</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.1</td>\n",
       "      <td>0.3</td>\n",
       "      <td>0.4</td>\n",
       "      <td>0.2</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>3</th>\n",
       "      <td>videos/GPT4_themes/fashion/fabio_fashion_18.mp4</td>\n",
       "      <td>fashion</td>\n",
       "      <td>fabio_fashion_18</td>\n",
       "      <td>1</td>\n",
       "      <td>14</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.8</td>\n",
       "      <td>0.1</td>\n",
       "      <td>0.1</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.1</td>\n",
       "      <td>0.4</td>\n",
       "      <td>0.3</td>\n",
       "      <td>0.2</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>4</th>\n",
       "      <td>videos/GPT4_themes/fashion/fabio_fashion_19.mp4</td>\n",
       "      <td>fashion</td>\n",
       "      <td>fabio_fashion_19</td>\n",
       "      <td>1</td>\n",
       "      <td>14</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.8</td>\n",
       "      <td>0.1</td>\n",
       "      <td>0.1</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.1</td>\n",
       "      <td>0.4</td>\n",
       "      <td>0.4</td>\n",
       "      <td>0.1</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "</div>"
      ],
      "text/plain": [
       "                                                path    theme  \\\n",
       "0     videos/GPT4_themes/fashion/fabio_fashion_3.mp4  fashion   \n",
       "1  videos/GPT4_themes/fashion/chamillioneyes_fash...  fashion   \n",
       "2     videos/GPT4_themes/fashion/fabio_fashion_2.mp4  fashion   \n",
       "3    videos/GPT4_themes/fashion/fabio_fashion_18.mp4  fashion   \n",
       "4    videos/GPT4_themes/fashion/fabio_fashion_19.mp4  fashion   \n",
       "\n",
       "                      video  tries  extracted_probs nicotine vaping  \\\n",
       "0           fabio_fashion_3      1               14            0.05   \n",
       "1  chamillioneyes_fashion_1      1               14             0.0   \n",
       "2           fabio_fashion_2      1               14             0.0   \n",
       "3          fabio_fashion_18      1               14             0.0   \n",
       "4          fabio_fashion_19      1               14             0.0   \n",
       "\n",
       "  cannabis vaping fashion entertainment lifestyle age_0-11 age_12-17  \\\n",
       "0            0.05     0.8           0.1       0.1      0.0      0.05   \n",
       "1             0.0     0.8           0.2       0.1      0.0       0.0   \n",
       "2             0.0     0.8           0.1       0.1      0.0       0.0   \n",
       "3             0.0     0.8           0.1       0.1      0.0       0.0   \n",
       "4             0.0     0.8           0.1       0.1      0.0       0.0   \n",
       "\n",
       "  age_18-21 age_22-25 age_26-30 age_31-40 age_41-50 age_51-60 age_61_plus  \n",
       "0       0.2       0.4      0.25       0.1       0.0       0.0         0.0  \n",
       "1       0.2       0.5       0.2       0.1       0.0       0.0         0.0  \n",
       "2       0.1       0.3       0.4       0.2       0.0       0.0         0.0  \n",
       "3       0.1       0.4       0.3       0.2       0.0       0.0         0.0  \n",
       "4       0.1       0.4       0.4       0.1       0.0       0.0         0.0  "
      ]
     },
     "execution_count": 346,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "df2 = df.merge(result_df, on = 'video')\n",
    "df2.head()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 364,
   "id": "711f1155-e3cf-4645-acd1-0cafe063b352",
   "metadata": {},
   "outputs": [],
   "source": [
    "df2.to_parquet(\"data/gpt-video-probabilites.parquet\")"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "6a0a4578-560c-4b93-aa95-2cb0ceaa3777",
   "metadata": {},
   "source": [
    "## Evaluate accuracy and AUC"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 347,
   "id": "7a861cc2-5ae8-4673-8a90-09b34f097076",
   "metadata": {},
   "outputs": [],
   "source": [
    "import sklearn, sklearn.metrics"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 348,
   "id": "9b8ca62b-a589-4fdf-9160-7a157c78ffe2",
   "metadata": {},
   "outputs": [],
   "source": [
    "# Via https://gist.github.com/doraneko94/e24643136cfb8baf03ef8a314ab9615c\n",
    "# Formula based on these articles:\n",
    "# Hanley and McNeil, The meaning and use of the area under a receiver operating characteristic (ROC) curve. Radiology (1982) 43 (1) pp. 29-36.\n",
    "# Fogarty, Baker and Hudson, Case Studies in the use of ROC Curve Analysis for Sensor-Based Estimates in Human Computer Interaction, Proceedings of Graphics Interface (2005) pp. 129-136.\n",
    "\n",
    "from sklearn.metrics import roc_auc_score\n",
    "from math import sqrt\n",
    "\n",
    "def roc_auc_ci(y_true, y_score, positive=1):\n",
    "    AUC = roc_auc_score(y_true, y_score)\n",
    "    N1 = sum(y_true == positive)\n",
    "    N2 = sum(y_true != positive)\n",
    "    Q1 = AUC / (2 - AUC)\n",
    "    Q2 = 2*AUC**2 / (1 + AUC)\n",
    "    SE_AUC = sqrt((AUC*(1 - AUC) + (N1 - 1)*(Q1 - AUC**2) + (N2 - 1)*(Q2 - AUC**2)) / (N1*N2))\n",
    "    lower = AUC - 1.96*SE_AUC\n",
    "    upper = AUC + 1.96*SE_AUC\n",
    "    if lower < 0:\n",
    "        lower = 0\n",
    "    if upper > 1:\n",
    "        upper = 1\n",
    "    return (lower, upper)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 362,
   "id": "12cc37f7-d357-4409-8b79-8c1833f83c2b",
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "\n",
      "Analyzing fashion (prompt code: fashion)\n",
      "Theme prevalence (original df): 0.1961\n",
      "Theme prevalence (analysis df): 0.1961\n",
      "Analyzed sample size: 102\n",
      "AUC: 0.9951 0.9734-1.0000\n",
      "Accuracy: 0.9902\n",
      "\n",
      "Analyzing health (prompt code: lifestyle)\n",
      "Theme prevalence (original df): 0.1961\n",
      "Theme prevalence (analysis df): 0.1961\n",
      "Analyzed sample size: 102\n",
      "AUC: 0.9948 0.9724-1.0000\n",
      "Accuracy: 0.9608\n",
      "\n",
      "Analyzing ecigs (prompt code: nicotine vaping)\n",
      "Theme prevalence (original df): 0.2157\n",
      "Theme prevalence (analysis df): 0.2157\n",
      "Analyzed sample size: 102\n",
      "AUC: 0.8847 0.7900-0.9793\n",
      "Accuracy: 0.8235\n",
      "\n",
      "Analyzing marijuana (prompt code: cannabis vaping)\n",
      "Theme prevalence (original df): 0.1961\n",
      "Theme prevalence (analysis df): 0.1961\n",
      "Analyzed sample size: 102\n",
      "AUC: 0.9875 0.9529-1.0000\n",
      "Accuracy: 0.9608\n",
      "\n",
      "Analyzing technology (prompt code: entertainment)\n",
      "Theme prevalence (original df): 0.1961\n",
      "Theme prevalence (analysis df): 0.1961\n",
      "Analyzed sample size: 102\n",
      "AUC: 1.0000 1.0000-1.0000\n",
      "Accuracy: 0.9608\n"
     ]
    }
   ],
   "source": [
    "themes = df2.theme.unique()\n",
    "theme_cols = {'fashion': 'fashion',\n",
    "              'health': 'lifestyle',\n",
    "             'ecigs': 'nicotine vaping',\n",
    "             'marijuana': 'cannabis vaping',\n",
    "             'technology': 'entertainment'}\n",
    "\n",
    "for theme, theme_col in theme_cols.items():\n",
    "    print(f\"\\nAnalyzing {theme} (prompt code: {theme_col})\")\n",
    "    true_val = (df2['theme'] == theme).astype(int)\n",
    "    print(f\"Theme prevalence (original df): {np.mean(true_val):0.4f}\")\n",
    "    predicted_val = df2[theme_col].astype(float)\n",
    "    analysis_df = pd.concat([true_val, predicted_val], axis = 1)\n",
    "    # Drop rows missing the predicted prob or the true label (hopefully none are missing the true label).\n",
    "    analysis_df.dropna(inplace = True)\n",
    "    print(f\"Theme prevalence (analysis df): {np.mean(analysis_df['theme'].values):0.4f}\")\n",
    "    print(\"Analyzed sample size:\", analysis_df.shape[0])\n",
    "    roc_result = sklearn.metrics.roc_auc_score(analysis_df['theme'], analysis_df[theme_col])\n",
    "    auc_ci = roc_auc_ci(analysis_df['theme'], analysis_df[theme_col])\n",
    "    print(f\"AUC: {roc_result:0.4f} {auc_ci[0]:0.4f}-{auc_ci[1]:0.4f}\")\n",
    "    accuracy = np.mean((predicted_val > 0.5).astype(int) == true_val)\n",
    "    print(f\"Accuracy: {accuracy:.4f}\")"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "id": "df6dc799-2a21-4f1a-a5c2-09fdd6ed87db",
   "metadata": {},
   "outputs": [],
   "source": []
  }
 ],
 "metadata": {
  "kernelspec": {
   "display_name": "Python 3 (ipykernel)",
   "language": "python",
   "name": "python3"
  },
  "language_info": {
   "codemirror_mode": {
    "name": "ipython",
    "version": 3
   },
   "file_extension": ".py",
   "mimetype": "text/x-python",
   "name": "python",
   "nbconvert_exporter": "python",
   "pygments_lexer": "ipython3",
   "version": "3.12.8"
  }
 },
 "nbformat": 4,
 "nbformat_minor": 5
}