Mapeamento dos preços do mercado imobiliário obtidos a partir de web scraping de páginas de anúncios dos imóveis urbanos

Abstract

Due to the characteristics that compose real estate speculation in large cities and the way it currently operates, the prices of real estate developments reflect and intensify the price of urban land. The present work mapped the spatial concentration of real estate market prices obtained through the search of online classifieds for the sale of properties classified by type and spatialized by neighborhood, applied in the city of Salvador-Bahia, as a case study. The method proposes the use of data available on the web recovered through the online data scraping technique (web scraping), through the use of extensions installed in the browser from a Hypertext Markup Language (HTML) website for real estate marketing. For this, it presents the debugging of the database based on exclusion criteria for the inconsistencies associated with the web scraping database. In the approach, attention is found to the potential and data of big data. For spatial analysis, cluster and outlier analysis methods are used, such as the Location Indicator Spatial Autocorrelation (LISA), which is decomposed by the Moran Global Index. As a result of the research, it was observed that the Olx base presented lower value (completeness), lower volume compared to Imovelweb, but greater variety regarding the spatial coverage of properties by neighborhood. In this context, the mapping of the distribution of average prices per square meter (m²) showed the concentration of high prices in the upscale neighborhoods located on the Atlantic coast of Salvador, as well as low prices in the popular neighborhoods of the core and suburb of Baía de Todos os Santos