Madagascar energy insights: a multi-layered ground-truth and geospatial dataset for AI-powered solar forecasting and energy planning
PI and co-PIs: Fabienne Rafidiharinirina (MAIDI, Madagascar); Mamisoa Ramanitrarivo (MAIDI, Madagascar); Fabrice Ramamonjy (MAIDI, Madagascar); Mamy Rakotoarisoa (CNIA Madagascar); Fandresena Rajerison (Solar Madagascar Power, Madagascar)
Funding amount: $133,216.30
Project overview: To effectively address both climate change and sustainable development goals, it is important that strategies for deploying clean energy consider not merely the extent of deployments but also the impacts on energy access, reliability, and usage. In low- and middle-income countries with high rates of power shortages, the adoption of off-grid solar systems can have implications for both decarbonization and energy security; assessing these outcomes in practice can enable better data-driven planning of both on-grid and off-grid energy solutions. However, existing datasets often lack the comprehensiveness, resolution, and on-the-ground survey data necessary to properly assess these outcomes. To answer this, this project aims to build a dataset that integrates solar panel locations and specifications, climate and weather data, data on grid power production/consumption and power shortages, and household-level survey data. Collected via national strategic partnerships in Madagascar, this dataset will help national energy stakeholders to better plan energy distribution at a city level. In particular, such a dataset will enable the assessment of current solar panel reliability under different climate conditions, help identify areas with unmet energy need and low solar potential, and thereby enable better forecasting and integrated energy planning that considers both decarbonization and energy reliability needs.
Power & Energy Cities & Urban Planning Computer Vision & Remote Sensing Time-series Analysis