We worked on this project with my team during the Smart India Hackathon 2022, which involved prediction of bias-corrected temperature from the Numerical Weather Prediction (NWP) model, and visualising the same through a web application.

We observed approximately 100% correction in bias for short durations, and approximately a 30 times improvement in RMSE.

I worked on developing a Flask application to visualize original and corrected values along with bias and error for any location in India, for a duration of 6 months.

Implementation

Check out the project presentation here, and the GitHub repository for the project here.