A.U.R.A

Analyze. Understand. Relax. Adapt

Basic Details

Team Name: GirlCode

Team Members

Member 1 : Lana Anvar - DCS, CUSAT
Member 2 : S Sutharya - DCS, CUSAT
Member 3 : Lakshmikha Rejith - SOE, CUSAT

Hosted Project Link

Access our project here - https://aquamarine-salamander-a3704c.netlify.app/

Project Description

AURA is a comprehensive application designed to analyze and visualize the emotional and mental state of users based on their textual input. The application detects emotions, analyzes sentiment, and generates personalized advice to help users manage their stress levels.

The Problem Statement

In today's fast-paced world, individuals are increasingly experiencing high levels of stress and anxiety due to various personal and professional challenges. Despite the availability of numerous mental health resources, many people struggle to find personalized and immediate support that addresses their unique emotional states. Traditional methods of stress management often fail to provide real-time, tailored advice that can help individuals cope with their specific situations.

The Solution

The solution is to develop a web application that can accurately analyze a person's emotional and mental state based on their textual input and provide personalized, actionable advice to help them manage their stress levels effectively. This solution leverages advanced natural language processing (NLP) techniques to detect emotions, analyze sentiment, and generate contextually relevant advice, thereby offering a comprehensive and user-friendly tool for mental health support.

Technical Details

Technologies/Components Used

Languages Used : Python, JavaScript

Frameworks Used : React

Libraries Used :
Python - FastAPI, Pydantic, Transformers, Torch, Requests, Dotenv, Spacy, NLTK, Logging
JavaScript - React, Vite, ESLint

Models Used : bhadresh-savani/distilbert-base-uncased-emotion, GEMINI

Tools Used : Visual Studio Code, Git, GitHub, Uvicorn

Implementation

Installation

To set up the project, follow these steps:

1. Clone the repository:

   git clone https://github.com/yourusername/StressVisualizer.git
   cd StressVisualizer

2. Create and activate a virtual environment:

   python -m venv venv
   # On Windows
   .\venv\Scripts\activate
   # On macOS/Linux
   source venv/bin/activate

3. Install the required dependencies:

   pip install -r requirements.txt

Run

To run the FastAPI application, use the following command:

uvicorn main:app --reload

This will start the server, and you can access the API at http://127.0.0.1:8000.

Project Documentation

Screenshots

This is the image of the final python output obtained from the models after emotion analysis and text generation.

The image of our static frontend.

The final combined output.

Project Demo

Video

Watch the video here - https://drive.google.com/file/d/1x49GieXJ_4lWub7-4AEsvz7TyJSp-OLD/view?usp=sharing

Team Contributions

Lana Anvar : Worked mainly in the integration of the frontend and the backend.
S Sutharya : Designed and created the frontend.
Lakshmikha Rejith : Worked mainly in the model integration and optimization.

License

This project is licensed under the MIT License. See the LICENSE file for details.