This is my solution for the WiDS 2022 Datathon
To help combat climate change, this year's WiDS datathon was to predict building energy consumption using a dataset that details buildings characteristics and weather conditions of the area the buildings are at. My solution included minimal feature enginnering and used a blend of a LightGBM and a Random Forest model to get the final predictions. My solution's RMSE score of 22.573 on private test data ranked 81/829 in the datathon, you can view the leaderboard here.