This project builds three separate models; Ridge regression model, XGBoost regression model and Neural Network regression model to predict household and village level electricity consumption in Rwanda in 2019. The predictors used are obtained from three categories
- electricity reliability
- Proximity to socio-economic infrastructure
- Household and village level charaterisitics such as land cover changes and building rooftop size.
The dependent variable is annual electricity consumption in 2019. The objective of the project is to predict household and village level consumption and determine features important in making the predictions.