Comparative evaluation of classical statistical and machine learning models for monthly atmospheric humidity forecasting in Bangladesh

This repository contains the data and code required to reproduce the analyses presented in the study. It includes the aggregated monthly dataset, the Jupyter Notebook containing the complete analysis workflow, Bangladesh administrative boundary shapefiles (SHP), the country mask used for spatial data extraction, and a README file describing the repository contents, data sources, software requirements, and instructions for reproducing the results. The dataset covers January 1981 to December 2025 and includes monthly relative humidity (RH2M), temperature (T2M), surface pressure (PS), and monthly total corrected precipitation (PRECTOTCORR_SUM), represented as reanalysis-derived estimates from the MERRA-2 system.