InsightAI

DataDriven Insights for Patients and Healthcare Professionals

Overview

InsightAI is a comprehensive health analytics platform that helps healthcare professionals analyze patient physical activity data. The application features AI-powered insights, customizable visualizations, and an interactive chat interface for querying patient data.

Features

• Physical activity and sedentary behavior analysis
• WHO recommendation progress tracking
• AI-powered insights using OpenAI and DeepSeek models
• Interactive chat interface with natural language data querying
• Vizro AI integration for generating visualizations from natural language
• Exercise planning and management

Prerequisites

• Python 3.8 or higher
• OpenAI API Key
  • Get your API key from OpenAI Platform
  • Follow the OpenAI API Keys Guide for setup instructions
• (Optional) DeepSeek Running Locally on LMStudio
  • Download LM Studio
  • Follow the Quick Start Guide to set up DeepSeek locally

Installation

Step 1: Clone the repository

git clone <repository-url>
cd insightAI

Step 2: Create a virtual environment (recommended)

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

Step 3: Install dependencies

# Core dependencies
pip install dash dash-daq
pip install plotly pandas numpy
pip install python-dotenv requests

# LLM API dependencies (for OpenAI and DeepSeek)
pip install openai deepseek-ai

# Vizro and Vizro AI for visualizations
pip install vizro vizro-ai

Step 4: Set up environment variables

1. Create a .env file in the root directory
2. Add the following variables:

OPENAI_API_KEY="your_openai_api_key"

Step 5: Run the application

python main.py

The application will start and be available at http://127.0.0.1:8050/ in your web browser.

Usage Guide

Data Analysis

• Select a participant from the dropdown menu
• Choose a date range and timeframe
• Toggle AI analysis on/off and select which models to use (OpenAI, DeepSeek)
• View physical activity and sedentary behavior charts
• Expand/collapse sections as needed

Chat Interface

• Use the "Chat with Data" section to ask questions about the patient's health data
• Ask natural language questions like:
  • "How does this patient's activity level compare to WHO recommendations?"
  • "Visualize activity trends over the selected time period"
  • "What's the weekly MPA and VPA distribution for this patient?"
  • "Show me the correlation between sedentary time and physical activity"
• The chat interface supports visualization generation through Vizro AI

Date Formats

The application handles European date format (DD.MM.YYYY) for data import and display.

Troubleshooting

• If you encounter API key issues, ensure your .env file is properly configured
• For visualization errors, check that Vizro and Vizro AI are correctly installed
• If DeepSeek model is not working, verify your local LM Studio setup

License

This project is licensed under the Apache License, Version 2.0 with the Commons Clause License Condition v1.0 - see the LICENSE file for details.

The Commons Clause means you may use, modify and redistribute the software freely, but you may not sell it - that is, you may not offer a product or service for a fee whose value derives substantially from the functionality of this software (including hosting or support services for it).