How aminilay/crop-yield-prediction-system works
This is an Indian crop-yield prediction product: a farmer supplies crop, location, season, year, area, and optional farm-specific inputs; the system assembles weather and agricultural features, predicts yield, and presents results through a dashboard or REST API. The graph also preserves the training pipeline, forecast and risk capabilities, and explainability. Source excerpts establish several data and inference relationships; other product capabilities are included where documented in the README, without assuming unsampled implementation details.
Main entry points: predict_pipeline.py Prediction Pipeline · app.py Streamlit Dashboard · data_transformation.py Data Transformation · weather_data.py Historical Weather
Components 20
User Interfaces
- Streamlit Dashboard
dashboard/app.py - REST API
api/main.py
Prediction Runtime
- Prediction Pipeline
src/pipeline/predict_pipeline.py - Yield Model
models/model.json - Feature Preprocessor
models/preprocessor.pkl - SHAP Explainability
dashboard/app.py - Weather Risk Assessment
src/components/weather_risk.py
Agricultural Features
- Historical Weather
src/components/weather_data.py - Seasonal Forecast
src/components/forecast_data.py - State Coordinates
src/utils/geo_coords.py - Irrigation Data
src/components/irrigation_data.py - Fertilizer Data
src/components/fertilizer_data.py - Agricultural Datasets
data/raw
Model Training
- Training Pipeline
src/pipeline/train_pipeline.py - Data Ingestion
src/components/data_ingestion.py - Data Transformation
src/components/data_transformation.py - Model Trainer
src/components/model_trainer.py
Shared Utilities
- Model Utilities
src/utils/common.py
Other components
- Farmer
- Weather Services
Connections 25
- Farmer → Streamlit Dashboard: enters inputs
README.md - Farmer → REST API: submits request
README.md - Streamlit Dashboard → Prediction Pipeline: requests prediction
- REST API → Prediction Pipeline: requests prediction
- Prediction Pipeline → Historical Weather: fetches weather
src/pipeline/predict_pipeline.py - Prediction Pipeline → Irrigation Data: gets irrigation
src/pipeline/predict_pipeline.py - Prediction Pipeline → Fertilizer Data: gets NPK
src/pipeline/predict_pipeline.py - Historical Weather → State Coordinates: looks up coordinates
src/components/weather_data.py - Seasonal Forecast → State Coordinates: looks up coordinates
src/components/forecast_data.py - Historical Weather → Weather Services: queries NASA POWER
README.md - Seasonal Forecast → Weather Services: queries Open-Meteo
README.md - Prediction Pipeline → Feature Preprocessor: transforms features
- Prediction Pipeline → Yield Model: predicts yield
- Streamlit Dashboard → SHAP Explainability: presents explanations
README.md - REST API → Weather Risk Assessment: serves risk results
README.md - Streamlit Dashboard → Weather Risk Assessment: shows risk alerts
README.md - Training Pipeline → Data Ingestion: runs ingestion
src/pipeline/train_pipeline.py - Training Pipeline → Data Transformation: runs transformation
src/pipeline/train_pipeline.py - Training Pipeline → Model Trainer: runs training
src/pipeline/train_pipeline.py - Data Transformation → Historical Weather: enriches with weather
src/components/data_transformation.py - Data Transformation → Irrigation Data: enriches irrigation
src/components/data_transformation.py - Data Transformation → Fertilizer Data: enriches fertilizer
src/components/data_transformation.py - Model Trainer → Yield Model: produces model
- Model Utilities → Yield Model: loads model
src/utils/common.py - Data Ingestion → Agricultural Datasets: reads datasets