Precificação de imóveis na cidade de Belo Horizonte: uma análise preditiva usando regressão linear e random forest

Abstract

The real estate market plays a significant role in the economy of Brazilian municipalities. This study aims to apply predictive models for pricing residential properties, such as houses and apartments, in Belo Horizonte, Minas Gerais, Brazil. Using a public database related to the collection of the Buildings Transfer Tax (ITBI), available on the open data portal of the Belo Horizonte city hall, an analysis of these data was conducted using two techniques: linear regression and Random Forest. Variables such as the city's region, property finishing type, and building age were considered as potential predictors. The results indicated that the Random Forest model achieved 82% accuracy in predicting property prices in Belo Horizonte based on these factors. This research contributes to illustrating how statistical methods can be used to guide and assess a more realistic property valuation, aligning with market prices in the municipality's real estate sector.