Tempest00-max/nigeria-gnss-meteorology

Nigeria GNSS Meteorology System — ML-based Tm model for 37 stations # NiMet-GNSS | Nigeria GNSS Meteorology System **Autonomous GNSS-Meteorological Analysis for Regional PWV Estimation in Nigeria** ## Features - 37 stations across all 36 Nigerian states + FCT - Machine Learning-based Weighted Mean Temperature (Tm) model - Real-time ZTD, ZWD, ZTD, and PWV computation - Model comparison: Nigeria vs Bevis vs GPT3 vs 6 baselines - Interactive dashboard with Leaflet mapping ## Models - Linear Regression (RMSE: 2.015 K) - Random Forest (RMSE: 2.052 K) - XGBoost (RMSE: 2.026 K) ## Validation - R² = 0.9008 - 90.5% improvement over Bevis global model - 6-hourly temporal resolution, 2017-2022 coverage ## API - POST /api/v1/tm/compute - POST /api/v1/ztd/compute - POST /api/v1/tm/compare - POST /api/v1/timeseries - POST /api/v1/national/summary ## Run Locally ```bash pip install -r requirements.txt python backend_v3.py # Open 4_frontend_v3.html in browser Author Gbadamosi Tolulope | Geodesy MSc | Nigeria