CS-245 Machine Learning Project - Fall 2025
Agri-Intel is a machine learning-based decision support system designed to optimize wheat production in Punjab. By integrating Government Agricultural Statistics, MODIS Satellite Data (NDVI), and NASA POWER Climate Data, this system predicts crop yields and recommends optimal fertilizer usage for sustainable agriculture (Theme T4).
app.py: The main Streamlit application (Proof of Concept Dashboard).final_model.pkl: The trained Gradient Boosting Machine Learning model.final_dataset_ml_ready.csv: The merged master dataset used for predictions.files/punjab_districts_cleaned.geojson: Spatial data for the interactive map.Data_pipeline.ipynb: Notebook containing data collection and processing scripts.requirements.txt: List of Python dependencies.
- Ensure you have Python installed (version 3.8 or higher).
- Install the required libraries using pip:
pip install -r requirements.txt
Open your terminal in the project directory.
Run the following command:
Bash
streamlit run app.py
The dashboard will automatically open in your browser at http://localhost:8501.
or visit
https://asjad2401-agri-intel-streamlitapp-apuzxg.streamlit.app/
Select District: Choose a district from the sidebar to load its historical soil and climate defaults.
Analyze Yield: Adjust the sliders (Fertilizer, NDVI, Rainfall) and click "Analyze Yield" to see the prediction.
Optimization Simulation: Scroll down to the "What-If" section to see the AI's recommendation for maximizing yield based on your inputs.
Developed by: Muhammad Asjad