Data Mines in Real Estate Web Pages: Investigation of Changes in the Czech Real Estate Market Based on Elasticity and on Modified Price Volume Indicator.

A vast number of advertisements on the websites of real estate companies form deep unexplored data mines. Real estate data can be obtained daily by parsing such websites automatically. Specifically, data on the area of apartments and its price are commonly available. Thus, the real estate market could be inspected in each municipality of the Czech Republic. We have developed our own software tool for obtaining this data; it collected data during the period between June 2019 and March 2021. In addition to the price, we also monitored the number of advertisements in each municipality daily. The aim of this study is to propose a methodology for comparing the development of the real estate market in different districts. The comparison of regions is based on cluster analysis. In our study, each region was represented by 29 monthly averages of the number of advertisements and 29 monthly averages of average apartment prices per 1 m2. However, we performed clustering on derived variables: the monthly value of elasticity and a modified price-volume indicator. The obtained clusters are graphically represented. In addition, the structure of market changes over time is economically interpreted.