StudyBuddy: An Intelligent Education App

StudyBuddy is an innovative educational application designed to streamline classroom management and enhance student engagement through intelligent digital interactions. It offers a range of features including class creation, FAQ management, file uploads for teachers, and interactive class joining and FAQ interactions for students.

Important

We host our app on the free tier of Azure app services, so it is SLOW. Please be patient when waiting for responses from the chatbot.

Demo the app

User accounts

• Sample teacher account (Username: t1, Password t1)

• Sample student accounts (Username s1, Password s1)

Try out some of the following prompts:

Sample Teacher prompts:

• Log in as t1
• Select the REL 3367 class (one of our actual university courses!) from the dropdown menu
• Select the Early Islam Module from the dropdown menu
• Ask Study Buddy: Please make me an assignment that will allow students to learn about the characteristics of the Early Islamic time period and issues that arose following the death of Prophet Muhammad. This assignment should be a medium-hard level of difficulty and require some critical thinking components.

Warning

Please reset the chat before selecting a new class, and do not interact with the sidebar while the chat first loads to avoid errors.

or try:

• Log in as t1
• Select BUS500
• Select all modules
• Ask Study Buddy: Can you tell me which students submitted notes, and tell me who took detailed notes and who didn't?

Sample Student prompt:

• Log in as s1
• Select the BUS 5000 class from the dropdown menu
• Select the Introduction module from the dropdown menu
• Ask Study Buddy: Can you explain to me the purpose and value of the education and skill set that comes from an MBA? Please give me some exercises I can work on throughout this course based on the textbook to practice the skills I will learn. Make some of the exercises easy and some more difficult.

How it Works

StudyBuddy operates through a user-friendly interface, allowing teachers to create classes, modules within their classes, manage FAQs, and upload files.

Teachers provide descriptions and key learning outcomes for every class and module within the class, which the chatbot takes into account when answering questions to tailor its responses to the specifics of that class.

Students can join classes using a class code and interact with the class materials provided by the teacher. Both teachers and students can interface with StudyBuddy’s built in chatbot, which utilizes OpenAI’s Chatgpt 4 model to answer questions based on the specific class and module selected.

When a conversation begins, the user selects the modules they want to chat about. Once these are selected, the app pulls the relevant files from Azure blob storage, and the relevant contextual information for the class and each selected module from Azure SQL DB, and uploads this to the OpenAI API endpoint. The OpenAI assistant retrieves this information as necessary to inform its responses.

Teachers can use the chatbot to ask about the extent of their students’ knowledge, since the chatbot analyzes notes uploaded by students. They can also use it as a tool for lesson planning and ensuring they meet federal, state, and local education requirements. Students can use the chatbot to ask about anything pertaining to the class; the chatbot’s prompting is set up in a way in which it does not directly give students all the answers, rather it aims to act as a teacher in guiding them to the answer so the student actively learns in the process.

Features:

• Add Classes

• Add Modules within Classes

• Add/Answer FAQs

• Manage Assignments (Add/complete tasks)

• Interactive chatbot

Tech Stack

Azure Services

• Azure SQL Database

  • storing and retrieving user data

• Azure Storage

  • for file storage

• Azure App Service

  • to host the project as a web app with CI/CD

OpenAI Services

• OpenAI custom configured assistant

• Code interpreter and knowledge retrieval are enabled

Prerequisites to run locally

• Access to an Open API Key
• An Open AI Assistant
• Access to the above mention Azure and OpenAI services

Create an Azure account here if you do not already have one

Local Deployment

1. Clone the repository git clone in command line
2. Run pip install -r requirements.txt in your terminal to install necessary packages.
3. Configure a .env file in the root directory with the following variables:

OPENAI_API_KEY= Your API key for Open AI services  
OPENAI_ASSISTANT= Your OpenAI model
AZURE_SERVER= The server address for your Azure SQL Database  
AZURE_DATABASE= The name of your Azure SQL database  
AZURE_USERNAME= Your username for Azure SQL Database  
AZURE_PASSWORD= Your password for Azure SQL Database
AZURE_STORAGE_CONNECTION_STRING= Your connection for Azure Storage account  
AZURE_CONTAINER= Your name for Azure Blob storage container  

4. In command line, run streamlit run app.py

Production improvements

Deploying this to production, we would opt to use Azure OpenAI services as opposed to a regular OpenAI endpoint. Azure OpenAI services could easily replace OpenAI’s services in the framework we built, we just need access to an Azure Open AI key.

Built With

Streamlit - The framework used to build the app.