Real Estate Indices: Polish Housing Market Analysis
The indices are constructed using a database containing over 6 million housing listings across 28 Polish cities from 1996 to the present. These cities represent a significant portion (52%) of secondary market transactions in 2019.
Data collection involved multiple sources - earlier records came from digitized newspaper advertisements and local periodicals, while more recent data (from 2009 onward) was gathered through web-scraping of online real estate portals. View my indices for Poland, city prices, or rental markets.
Data Collection & Processing
Historical data requires digital conversion, and information availability varies depending on publication date and source. While basic details like price, location, and size were commonly available, older listings often omitted prices but included unusual information such as telephone availability in the apartment.
More recent listings from online portals provided additional information including:
- Precise location (district, housing estate, street)
- Asking price
- Floor
- Ownership type
- Usable area
- Construction technology
- Parking facilities
- Apartment quality
Database Development
Creating a uniform database required addressing several challenges. The database development process involved careful standardization of variables from diverse sources and establishing consistent location identifiers, which ranged from general city/district classifications to specific street addresses.
- Standardizing variables from diverse sources
- Establishing consistent location identifiers (ranging from general city/district to specific street addresses)
- Removing duplicate listings (same apartment advertised across multiple months or by different agencies)
- Eliminating outliers using Cook distance methodology
After thorough data cleaning, the final dataset contained 4 million unique property listings. These data served as the foundation for national and city-level indices.
Hedonic Index Construction
The hedonic House Price Index (HPI) for each city was developed using the rolling time-dummy method. The variables included in the models differed based on data source:
- District/estate
- Usable area
- Age of the building
- Construction technology
- Apartment quality
- District/estate
- Usable area
The research produced quarterly hedonic HPI values for selected Polish cities from Q1 1996. A national index for Poland was subsequently constructed using transaction share data from the Central Statistical Office (CSO) for 2019. Separate indices for rental markets across major cities are also available.