On Representativeness of Internet Data Sources for Real Estate Market in Poland

Abstract

Shifting paradigms in Official Statistics lead to the widespread use of administrative records in an effort to support or create an alternative for censuses and surveys. At the same time the demand for diversified detailed information is increasing. In order to meet this demand Official Statistics needs to seek new data sources. Internet data sources (IDS), or more generally, big data could be one of them. The potential usefulness of these new sources of statistical information should not be neglected.The aim of the paper is to report on a study intended to assess the representativeness of IDS for the real estate market in Poland. These sources could be used for describing the demand and supply on the secondary real estate market in a more detailed way than is possible with the existing methodology. The degree of representativeness is assessed on the basis of information from official surveys and other data sources. Due to the shortage of relevant literature on the subject, the article provides a definition of IDS and draws on insights from a study conducted by the author to enhance information from Official Statistics. The study involved using information on street names from the National Official Register of the Territorial Division of the Country (TERYT) to harmonize street names obtained from IDS. A special program for automated data collection (web spider ) was developed. All the calculations were made with R (R Core Team 2014) statistical software and additional R packages (XML, RCurl, httr and ggplot2).

A Two‐Step Procedure to Measure Representativeness of Internet Data Sources

“…In the light of the previous discussion, the following definition of an IDS is proposed, which extends the definition proposed earlier in Berȩsewicz ():…”

A Fully Automated Adjustment of Ensemble Methods in Machine Learning for Modeling Complex Real Estate Systems

“…e Internet source also offers information on a variety of property and neighbourhood characteristics that are difficult to find from other sources [20]. However, this source of information has been little used despite the existence of applications developed with great success in both the real estate sector as well as other sectors [54].…”

Big data w statystyce publicznej – nadzieje, osiągnięcia, wyzwania i zagrożenia

“…Natomiast wraz z rozwojem projektu powstała firma PriceStats, której celem było pobieranie informacji o cenach ze sklepów internetowych z blisko 60 krajów, dostarczanie dziennych, miesięcznych i kwartalnych wskaźników inflacji oraz przeprowadzanie badań makroekonomicznych. Oprócz tego pojawiły się również pierwsze prace związane z wykorzystaniem danych do opisu zmian cen ubrań [Griffioen, de Haan, Willenborg 2014], produktów spożywczych [Swier 2015] czy rynku nieruchomości [Beręsewicz 2015] w kontekście statystyki publicznej.…”