Rene-Twizerimana/RWIMBOGO_MAIZE_GROUGHT_ANALYSIS_ANALYTICAL_MULTIPLE_REGRESSION_MODEL

This is the primary research made on RWIMBOGO sector, GATSIBO DISTRICT in Rwanda country analysis the drivers of maize yield loss due to the drought that happens in GATSIBO DISTRICT IN RWANDA **Drought Impact & Strategy Effectiveness Modeling for Maize Production (Rwimbogo, Rwanda)** This repository contains a complete data science pipeline that analyzes how drought affects maize yield and how farmers’ coping strategies and institutional support reduce these losses. The project includes: Real-world-inspired agricultural, rainfall, and survey data Data preprocessing & feature engineering Correlation analysis Two multiple regression models: Model 1: Drivers of maize yield loss Model 2: Drivers of strategy effectiveness Model diagnostics & visualizations The goal is to quantify climate vulnerability and measure the impact of adaptation strategies using interpretable econometric models. ** Problem Statement** Farmers in drought-prone areas face repeated crop losses. However, not all farmers are affected equally. Some adopt strategies (e.g., drought-resistant seeds, irrigation), and some receive support (inputs, training, credit). This project answers: What factors increase maize yield loss? What factors reduce maize yield loss? Which factors make coping strategies more effective? Modeling Approach Step 1: Data Preparation Multiple datasets were merged and transformed into quantitative indices: Feature Description Drought_Frequency_Score How often droughts occur Drought_Effects_Score Severity of drought impacts Strategy_Adoption_Score Level of coping strategy use Support_Access_Score Access to inputs, training, credit Strategy_Effectiveness_Score Perceived impact of strategies Yield_Loss_Percentage % maize yield loss (target) Step 2: Correlation Analysis Pearson correlation was used to: Explore relationships Check expected directions Identify important predictors Detect multicollinearity Step 3: Regression Models 🔹 Model 1: Yield Loss Drivers (Main Model) Predicts: Yield_Loss_Percentage Using: Drought_Frequency_Score Drought_Effects_Score Strategy_Adoption_Score Support_Access_Score This is the core model of the project. 🔹 …