Statihouse®: Technological development based on data science to statistically explore the real estate sector

Abstract

The real estate sector requires better use of data science and the massive data, to favor informed management and decision making. The present study aims to describe a system (Statihouse®) that statistically characterizes properties for sale and predicts their offer price (in real time or near), using massive data from the internet, data science and a web application. It exposes the development of the system, emphasizing its technical characteristics, processes and subprocesses, and visualization for the user. We present examples of visualization, providing results derived from a sample of 24,935 used houses, offered for sale in Colombia (period January 2017 - May 2017), which were contrasted with some referents as an exploratory mode. This system is novel, since there were no antecedents that execute automatically from the collection of web data, to the visualization (in real time or near), with a multitasking scope (descriptive, comparative, evolutionary, correlational and predictive) in the Real estate sector, with a focus on property. This system evidences a case of success of data science in the real estate sector and serves as a guide and stimulus for new developments.