Zulach/Hydrosense-Kenya

# HydroSense-Kenya: Smart Irrigation Decision Support Numerical-methods capstone for climate-aware irrigation scheduling in Kenya. This system combines data cleaning, root-finding, linear systems, water balance simulation, and optimization to recommend efficient irrigation schedules. ## Scientific Question Given weather and soil-sensor data, how can we model water availability, estimate water deficit, simulate future soil moisture, and recommend an efficient irrigation plan that minimizes water use without exposing crops to moisture stress? ## Key Features - **Data Quality Control**: Robust cleaning of sensor faults and weather anomalies. - **Numerical Methods Engine**: Manual implementation of bisection, Newton-Raphson, and secant methods for root-finding; trapezoidal and Simpson's rules for integration; and LU decomposition for linear systems. - **Simulation**: Soil-water balance modeling using Euler and RK4 methods. - **Uncertainty Analysis**: Monte Carlo simulation with 1,000+ scenarios. - **Optimization**: Gradient-descent-based irrigation scheduling to minimize water waste. ## Project Structure - `src/` — Core modules (numerical methods, simulation, cleaning, etc.) - `tests/` — Comprehensive pytest suite (~25 tests) - `notebooks/` — Leveled Jupyter notebooks for the intellectual arc - `data/` — Raw and processed CSV datasets - `reports/` — Final scientific report and presentation ## Setup ```bash pip install -r requirements.txt ``` ## Running Tests ```bash pytest tests/ ``` ## Authors Numerical Methods Masterclass Team Feb-May 2026