Automodeller demo.ipynb

{
 "cells": [
  {
   "cell_type": "code",
   "execution_count": 7,
   "id": "78aa4217-eedf-41d3-a7d3-57cbdfa2c2bf",
   "metadata": {
    "tags": []
   },
   "outputs": [],
   "source": [
    "import numpy as np\n",
    "import pandas as pd\n",
    "from sklearn.linear_model import LinearRegression\n",
    "from sklearn.metrics import r2_score\n",
    "\n",
    "np.random.seed(42)\n",
    "\n",
    "# Simulating media variables with positive impact\n",
    "media1 = np.random.normal(100, 20, 100)\n",
    "media2 = np.random.normal(200, 30, 100)\n",
    "media3 = np.random.normal(150, 25, 100)\n",
    "\n",
    "# Simulating random variables with no impact\n",
    "random1 = np.random.normal(0, 1, 100)\n",
    "random2 = np.random.normal(0, 1, 100)\n",
    "random3 = np.random.normal(0, 1, 100)\n",
    "\n",
    "# Simulating highly correlated media variables\n",
    "media4 = media1 * 1.5 + np.random.normal(0, 5, 100)\n",
    "\n",
    "# Simulating negative variables like price\n",
    "price = np.random.normal(10, 0.5, 100)\n",
    "\n",
    "# Simulating distribution\n",
    "distribution = np.random.normal(500, 100, 100)\n",
    "\n",
    "# Creating a DataFrame\n",
    "data = pd.DataFrame({\n",
    "    'media1': media1,\n",
    "    'media2': media2,\n",
    "    'media3': media3,\n",
    "    'random1': random1,\n",
    "    'random2': random2,\n",
    "    'random3': random3,\n",
    "    'media4': media4,\n",
    "    'price': price,\n",
    "    'distribution': distribution,\n",
    "})\n",
    "\n",
    "# Simulating sales based on the variables\n",
    "true_coefficients = np.array([0.5, 0.3, 0.2, 0, 0, 0, -1, 0.2])\n",
    "X = data[['media1', 'media2', 'media3', 'random1', 'random2', 'media4', 'price', 'distribution']]\n",
    "sales = X.values @ true_coefficients + np.random.normal(0, 4, 100)\n",
    "data['sales'] = sales\n",
    "\n",
    "# Fit the linear regression model and check the R2 score\n",
    "model = LinearRegression()\n",
    "model.fit(X, sales)\n",
    "predictions = model.predict(X)\n",
    "\n",
    "\n"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "2a91c75a-63e9-47cc-8d23-fb54828bdd6b",
   "metadata": {
    "tags": []
   },
   "source": [
    "# Simulate some MMM Data\n",
    "\n",
    "I've simulated data with highly correlated media variables and variables which are random noise"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 2,
   "id": "e2db7197-1f28-4bb2-83f0-8a3c4bcd2513",
   "metadata": {
    "tags": []
   },
   "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>media1</th>\n",
       "      <th>media2</th>\n",
       "      <th>media3</th>\n",
       "      <th>random1</th>\n",
       "      <th>random2</th>\n",
       "      <th>random3</th>\n",
       "      <th>media4</th>\n",
       "      <th>price</th>\n",
       "      <th>distribution</th>\n",
       "      <th>sales</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>0</th>\n",
       "      <td>109.934283</td>\n",
       "      <td>157.538878</td>\n",
       "      <td>158.944684</td>\n",
       "      <td>-0.828995</td>\n",
       "      <td>-1.594428</td>\n",
       "      <td>0.926178</td>\n",
       "      <td>168.686368</td>\n",
       "      <td>9.738638</td>\n",
       "      <td>593.828381</td>\n",
       "      <td>244.519473</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>1</th>\n",
       "      <td>97.234714</td>\n",
       "      <td>187.380640</td>\n",
       "      <td>164.019613</td>\n",
       "      <td>-0.560181</td>\n",
       "      <td>-0.599375</td>\n",
       "      <td>1.909417</td>\n",
       "      <td>141.241244</td>\n",
       "      <td>10.524505</td>\n",
       "      <td>448.395527</td>\n",
       "      <td>215.216717</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>2</th>\n",
       "      <td>112.953771</td>\n",
       "      <td>189.718565</td>\n",
       "      <td>177.076281</td>\n",
       "      <td>0.747294</td>\n",
       "      <td>0.005244</td>\n",
       "      <td>-1.398568</td>\n",
       "      <td>173.778686</td>\n",
       "      <td>9.647828</td>\n",
       "      <td>509.612078</td>\n",
       "      <td>241.197278</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>3</th>\n",
       "      <td>130.460597</td>\n",
       "      <td>175.931682</td>\n",
       "      <td>176.345051</td>\n",
       "      <td>0.610370</td>\n",
       "      <td>0.046981</td>\n",
       "      <td>0.562969</td>\n",
       "      <td>202.469085</td>\n",
       "      <td>9.295769</td>\n",
       "      <td>453.772471</td>\n",
       "      <td>239.851346</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>4</th>\n",
       "      <td>95.316933</td>\n",
       "      <td>195.161429</td>\n",
       "      <td>115.558266</td>\n",
       "      <td>-0.020902</td>\n",
       "      <td>-0.450065</td>\n",
       "      <td>-0.650643</td>\n",
       "      <td>145.042573</td>\n",
       "      <td>9.221685</td>\n",
       "      <td>456.550377</td>\n",
       "      <td>212.171334</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>...</th>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>95</th>\n",
       "      <td>70.729701</td>\n",
       "      <td>211.559521</td>\n",
       "      <td>132.677260</td>\n",
       "      <td>-0.469176</td>\n",
       "      <td>0.538910</td>\n",
       "      <td>-0.510016</td>\n",
       "      <td>104.546819</td>\n",
       "      <td>10.526576</td>\n",
       "      <td>564.870989</td>\n",
       "      <td>226.691379</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>96</th>\n",
       "      <td>105.922406</td>\n",
       "      <td>173.484277</td>\n",
       "      <td>172.489997</td>\n",
       "      <td>-1.713135</td>\n",
       "      <td>-1.037246</td>\n",
       "      <td>-0.269875</td>\n",
       "      <td>160.514273</td>\n",
       "      <td>9.980222</td>\n",
       "      <td>483.288192</td>\n",
       "      <td>233.372647</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>97</th>\n",
       "      <td>105.221105</td>\n",
       "      <td>204.611753</td>\n",
       "      <td>157.682488</td>\n",
       "      <td>1.353872</td>\n",
       "      <td>-0.190339</td>\n",
       "      <td>-0.978764</td>\n",
       "      <td>151.576090</td>\n",
       "      <td>10.340750</td>\n",
       "      <td>514.671369</td>\n",
       "      <td>240.687471</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>98</th>\n",
       "      <td>100.102269</td>\n",
       "      <td>201.746262</td>\n",
       "      <td>170.321553</td>\n",
       "      <td>-0.114540</td>\n",
       "      <td>-0.875618</td>\n",
       "      <td>-0.444293</td>\n",
       "      <td>154.773539</td>\n",
       "      <td>10.014159</td>\n",
       "      <td>620.650897</td>\n",
       "      <td>256.470628</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>99</th>\n",
       "      <td>95.308257</td>\n",
       "      <td>165.710891</td>\n",
       "      <td>165.740721</td>\n",
       "      <td>1.237816</td>\n",
       "      <td>-1.382800</td>\n",
       "      <td>0.377300</td>\n",
       "      <td>142.037875</td>\n",
       "      <td>10.014878</td>\n",
       "      <td>418.306433</td>\n",
       "      <td>206.452280</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "<p>100 rows × 10 columns</p>\n",
       "</div>"
      ],
      "text/plain": [
       "        media1      media2      media3   random1   random2   random3   \n",
       "0   109.934283  157.538878  158.944684 -0.828995 -1.594428  0.926178  \\\n",
       "1    97.234714  187.380640  164.019613 -0.560181 -0.599375  1.909417   \n",
       "2   112.953771  189.718565  177.076281  0.747294  0.005244 -1.398568   \n",
       "3   130.460597  175.931682  176.345051  0.610370  0.046981  0.562969   \n",
       "4    95.316933  195.161429  115.558266 -0.020902 -0.450065 -0.650643   \n",
       "..         ...         ...         ...       ...       ...       ...   \n",
       "95   70.729701  211.559521  132.677260 -0.469176  0.538910 -0.510016   \n",
       "96  105.922406  173.484277  172.489997 -1.713135 -1.037246 -0.269875   \n",
       "97  105.221105  204.611753  157.682488  1.353872 -0.190339 -0.978764   \n",
       "98  100.102269  201.746262  170.321553 -0.114540 -0.875618 -0.444293   \n",
       "99   95.308257  165.710891  165.740721  1.237816 -1.382800  0.377300   \n",
       "\n",
       "        media4      price  distribution       sales  \n",
       "0   168.686368   9.738638    593.828381  244.519473  \n",
       "1   141.241244  10.524505    448.395527  215.216717  \n",
       "2   173.778686   9.647828    509.612078  241.197278  \n",
       "3   202.469085   9.295769    453.772471  239.851346  \n",
       "4   145.042573   9.221685    456.550377  212.171334  \n",
       "..         ...        ...           ...         ...  \n",
       "95  104.546819  10.526576    564.870989  226.691379  \n",
       "96  160.514273   9.980222    483.288192  233.372647  \n",
       "97  151.576090  10.340750    514.671369  240.687471  \n",
       "98  154.773539  10.014159    620.650897  256.470628  \n",
       "99  142.037875  10.014878    418.306433  206.452280  \n",
       "\n",
       "[100 rows x 10 columns]"
      ]
     },
     "execution_count": 2,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "data"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "ed61589d-29ed-41fe-b955-1237b493cc78",
   "metadata": {},
   "source": [
    "# Import the module and set your openai key"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 3,
   "id": "d06089be-3b61-41a0-92f8-9df24b8481a4",
   "metadata": {},
   "outputs": [],
   "source": [
    "from autoMMM.automodeller import AutoModeller\n",
    "import openai\n",
    "openai.api_key = \"your key here\"\n"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "5a0f1bd1-bcd8-4143-80fa-547a84280e85",
   "metadata": {},
   "source": [
    "# Select independent & dependent variables and initial alpha for the ridge regression"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 14,
   "id": "3d5cc7e5-e905-4102-9ce5-5b72579a5e8c",
   "metadata": {
    "tags": []
   },
   "outputs": [],
   "source": [
    "# Select the initial independent variables\n",
    "independent_vars = ['media1', 'media2']\n",
    "# Set the dependent variable\n",
    "dependent_var = 'sales'\n",
    "# Set the initial alpha value for Ridge Regression\n",
    "initial_alpha = 1.0\n"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "18c7c8c0-512b-4ecd-8192-d589e7a0f768",
   "metadata": {},
   "source": [
    "# Set the minimum and maximum number of iterations \n",
    "**(Warning, each iteration is a prompt request to openai, therefore incurs a cost)**"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 15,
   "id": "3535c23b-ba92-4c69-9419-6d5adaf1746d",
   "metadata": {
    "tags": []
   },
   "outputs": [],
   "source": [
    "min_iterations = 1\n",
    "max_iterations = 10\n"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "324ba7bd-590a-4f56-a05e-9d8570b0ad44",
   "metadata": {},
   "source": [
    "# Create an instance of the AutoModeller class & run the automodeller\n"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 16,
   "id": "853e820d-6be3-42b2-8eb6-e0372f05d6d1",
   "metadata": {
    "tags": []
   },
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "variable test counts\n",
      "{'media1': 1, 'media2': 1, 'media3': 0, 'random1': 0, 'random2': 0, 'random3': 0, 'media4': 0, 'price': 0, 'distribution': 0, 'sales': 0}\n",
      "variable test counts\n",
      "{'media1': 2, 'media2': 2, 'media3': 1, 'random1': 0, 'random2': 0, 'random3': 0, 'media4': 1, 'price': 1, 'distribution': 1, 'sales': 0}\n",
      "model count:\n",
      "1\n",
      "variable test counts\n",
      "{'media1': 3, 'media2': 3, 'media3': 2, 'random1': 1, 'random2': 1, 'random3': 1, 'media4': 1, 'price': 2, 'distribution': 2, 'sales': 0}\n",
      "model count:\n",
      "2\n",
      "variable test counts\n",
      "{'media1': 4, 'media2': 4, 'media3': 3, 'random1': 1, 'random2': 1, 'random3': 1, 'media4': 2, 'price': 3, 'distribution': 3, 'sales': 0}\n",
      "model count:\n",
      "3\n",
      "variable test counts\n",
      "{'media1': 5, 'media2': 5, 'media3': 4, 'random1': 2, 'random2': 2, 'random3': 2, 'media4': 3, 'price': 4, 'distribution': 4, 'sales': 0}\n",
      "model count:\n",
      "4\n",
      "variable test counts\n",
      "{'media1': 6, 'media2': 6, 'media3': 5, 'random1': 2, 'random2': 3, 'random3': 3, 'media4': 4, 'price': 5, 'distribution': 5, 'sales': 0}\n",
      "model count:\n",
      "5\n",
      "variable test counts\n",
      "{'media1': 7, 'media2': 7, 'media3': 6, 'random1': 3, 'random2': 3, 'random3': 3, 'media4': 5, 'price': 6, 'distribution': 6, 'sales': 0}\n",
      "model count:\n",
      "6\n",
      "variable test counts\n",
      "{'media1': 8, 'media2': 8, 'media3': 7, 'random1': 3, 'random2': 4, 'random3': 3, 'media4': 6, 'price': 7, 'distribution': 7, 'sales': 0}\n",
      "model count:\n",
      "7\n",
      "variable test counts\n",
      "{'media1': 9, 'media2': 9, 'media3': 8, 'random1': 4, 'random2': 4, 'random3': 3, 'media4': 7, 'price': 8, 'distribution': 8, 'sales': 0}\n",
      "model count:\n",
      "8\n",
      "variable test counts\n",
      "{'media1': 10, 'media2': 10, 'media3': 9, 'random1': 4, 'random2': 5, 'random3': 4, 'media4': 8, 'price': 9, 'distribution': 9, 'sales': 0}\n",
      "model count:\n",
      "9\n",
      "variable test counts\n",
      "{'media1': 11, 'media2': 11, 'media3': 10, 'random1': 5, 'random2': 5, 'random3': 4, 'media4': 9, 'price': 10, 'distribution': 10, 'sales': 0}\n",
      "model count:\n",
      "10\n",
      "\n",
      "Final Model Variables:\n",
      "['media1', 'media2', 'media3', 'media4', 'price', 'distribution']\n",
      "\n",
      "Average Alpha over all iterations:\n",
      "1.3\n"
     ]
    }
   ],
   "source": [
    "automodeller = AutoModeller(data, dependent_var, independent_vars, initial_alpha, min_iterations, max_iterations)\n",
    "# Run the automodeller\n",
    "automodeller.run()\n"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "a4c7354f-19f4-4586-a9b6-d313f922bfbf",
   "metadata": {},
   "source": [
    "# Plot the variable test counts\n",
    "**Variable test counts give an indication of which variables the model has favoured\n"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 17,
   "id": "adb69324-2ba4-4048-9c69-b59d054c02f9",
   "metadata": {},
   "outputs": [
    {
     "data": {
      "image/png": "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",
      "text/plain": [
       "<Figure size 640x480 with 1 Axes>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    }
   ],
   "source": [
    "automodeller.plot_variable_test_counts()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "id": "63655ab3-72f2-4042-9a9e-a1ded0e3d621",
   "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": "3.11.0"
  }
 },
 "nbformat": 4,
 "nbformat_minor": 5
}