Our Mission: To empower communities, NGOs, and public health officials in Haiti with accessible, real-time, AI-driven data to ensure water safety and prevent waterborne diseases.

AquaLERT is an end-to-end water quality monitoring and analysis platform built for the #Hack4Haiti initiative. It combines machine learning, computer vision, and generative AI to provide instant, actionable insights on water potability, bridging the critical information gap that endangers millions.

In Haiti, over 4 million people lack access to a basic water source, and waterborne diseases remain a leading cause of mortality. The primary barriers to ensuring water safety are:

💰 Cost & Accessibility: Traditional laboratory testing is prohibitively expensive (50−200 per sample) and inaccessible to remote communities.

⏰ Time Delay: Lab results can take days or weeks, while contaminated water can cause illness within hours.

📊 Data Gap: Public health officials lack the real-time, high-level data needed to identify contamination hotspots, predict outbreaks, and allocate resources effectively.

AquaLERT tackles this challenge with a comprehensive, three-pillar solution:

ML-Powered Predictions: Uses a highly-tuned LightGBM model to predict water potability from low-cost sensor data with high accuracy.

Immediate Results: Provides a clear "Potable" or "Not Potable" verdict in under 30 seconds.

Offline Capability: Designed to work in remote areas with limited connectivity.

Google Gemini Integration: Translates complex data into clear, actionable public health advice.

Multilingual Support: Delivers recommendations in both English and Haitian Creole to ensure accessibility for all users.

Contextual Guidance: Provides specific instructions for treating water, securing sources, and long-term planning.

Live Interactive Map: Every test is geotagged and contributes to a live map, creating a crowdsourced water quality network.

Hotspot Detection: Empowers NGOs and health officials to identify high-risk areas and deploy resources efficiently.

Data-Driven Dashboards: Provides aggregated statistics and trend analysis for strategic planning and impact assessment.

AquaLERT follows a simple yet powerful four-step data flow:

Collect Data: A user in the field inputs sensor readings or uploads a photo of a water sample via the Streamlit frontend.

AI Analysis: The Flask backend receives the data.

For sensor data, the pre-trained LightGBM model makes a potability prediction.

For images, a computer vision model provides a preliminary visual assessment.

Get Results: The model's output is passed to the Google Gemini API, which generates a detailed, multilingual advisory. This complete report is sent back to the user.

Share Impact: The geotagged result is saved to a central database, instantly updating the Live Water Map and community dashboards.

Category Technologies Frontend Streamlit, Folium, Plotly, Pandas Backend Flask, Python ML & AI Scikit-learn, LightGBM, Prophet (for forecasting), Computer Vision (e.g., TensorFlow/PyTorch), Google Gemini API Database SQLite (for development), PostgreSQL (for production) Deployment Docker, Render / Heroku / AWS Elastic Beanstalk

Follow these instructions to set up and run the AquaLERT platform on your local machine.

Python 3.9 or higher

pip package manager

Git

Clone the repository:

git clone https://github.com/Saikumar1801/AquaLERT.git
cd AquaLERT

python -m venv venv
source venv/bin/activate  # On Windows, use `venv\Scripts\activate`

The project is split into a frontend and backend. Install requirements for both.

# Install requirements
pip install -r requirements.txt

You will need an API key for Google Gemini.

Create a file named .env inside the backend directory: backend/.env

Add your API key to this file:

GEMINI_API_KEY="YOUR_API_KEY_HERE"

Running the Application

You need to run two separate processes in two different terminals.

Run the Backend:

# In your first terminal
cd backend
python app.py

The backend server will start, typically on http://127.0.0.1:5000.

Run the Streamlit Frontend:

# In your second terminal
cd frontend
streamlit run Home.py

The Streamlit application will open in your web browser, usually at http://localhost:8501.

You can now interact with the full AquaLERT platform!

📂 Project Structure

aqualert/
├── backend/
│   ├── app.py                # Flask application logic and API endpoints
│   ├── models/
│   │   └── potability_model.pkl # Pre-trained LightGBM model
│   ├── .env                    # Environment variables (API keys)
│   └── requirements.txt        # Backend Python dependencies
│
├── frontend/
│   ├── Home.py                 # Main landing page for the Streamlit app
│   ├── pages/
│   │   ├── 1_Live_Water_Map.py # The interactive map page
│   │   ├── 2_Real-Time_Test.py # The sensor data input and analysis page
│   │   └── 3_Visual_Analysis.py# The image upload and analysis page
│   └── requirements.txt        # Frontend Python dependencies
│
├── notebooks/
│   ├── 1_Data_Exploration.ipynb # EDA on the water quality dataset
│   ├── 2_Model_Training.ipynb   # Final model training and evaluation
│   └── 3_Forecasting_POC.ipynb  # Prophet forecasting proof-of-concept
│
├── .gitignore
└── README.md

👥 Team & Acknowledgements

Saikumar - Project Lead & Backend/ML Developer

Saikumar - Frontend Developer

Saikumar - Data Scientist & UI/UX

We would like to thank the organizers of #Hack4Haiti for this incredible opportunity to work on a project with meaningful social impact. We also acknowledge the public datasets from Kaggle that made our initial model training possible.