This repository contains the whole codebase developed within the ING Lion's Den Modelling Challenge by the team of 4 Data Scientists, inc. myself, Mateusz Doliński, Anna Warno and Dominika Bankiewicz. We successfully accomplished 1st stage of the contest to finally get classified in TOP 6 best teams.
We were asked to developed predictive model for panel data, in order to improve the Bank's capital managment in the context of climate risk. In our analysis we implemented state-of-the-art univariate time series forecasting models inc. T-BATS, NBeats and Prophet as well as more traditional statistical models.