Machine Learning for Exploration of the Factors Affecting the Building Energy Usage: A Case Study

Abstract

This paper presents a thorough exploratory analysis of one of the data research tracks involving the building energy usage data of different cities of France. To gain a deep understanding of the dataset and the various attributes that comprise the data, we performed an exploratory data analysis and used data visualization to study the different variables and the correlation that exists between them. This was done to determine which factors really contribute to having a huge effect on the energy (gas) consumption leading to climate change. In our study, we also predicted the level of its consumption using a Decision-tree based algorithm and Light GBM, which was found to give satisfactory results.