Extensive analysis of rain in australia by exploratory data analysis, feature engineering and modeling

Abstract

Accurate rainfall forecasts help in planning outdoor activities, agricultural practices, and water resource management, thereby mitigating the impact of extreme weather events. This article provides an in-depth analysis of rainfall in Australia, focusing on predicting whether it will rain tomorrow using logistic regression. The research aims to develop an accurate model to help predict rainfall events for better preparedness and planning. We obtained datasets from a number of Australian weather stations. The dataset contains 142,193 daily weather observations spanning approximately ten years. The recorded information includes various details such as date, location, humidity, wind direction, clouds, temperature, etc. This shows that the model performs well in distinguishing between rainy and non-rainy days with an accuracy of about 0.875. The findings of this study have important implications for various stakeholders including meteorologists, disaster management agencies, and the public.