Training & validation set for coordinate-free UHI Index predictive model (EY Data Challenge 2025)

This dataset contains 2 CSV files. The final_training_data.csv file contains the cleaned data that was created by extracting satellite band data from Sentinel-2 and Landsat, and combining those with building and weather data. Two public Kaggle notebooks explain how the data was extracted and prepared and can be found here (Satellite Data Extraction (EY Data Challenge 2025)) and here (Data Preparation (EY Data Challenge 2025)). The dataset was used to train the model we used to predict UHI index values based on the EY Data Challenge 2025 and our solution is explained in this public Kaggle notebook (Predicting UHI Index (EY Data Challenge 2025)). The second CSV file included final_validation_data.csv was the external validation set we fed into our model and predicted the UHI index values. However, we did not have access to the true UHI index values since those were not available to us. The online submission system of the competition was the one grading the predictions. We have included that file for reference only.