Causal discovery on geospatial data: Investigating the impact of amenity proximity on residential property values

Shalaleh Hillebrand, Arjen Hommersom, Jesse Heyninck, Natasha Alechina

Track: Type A (Regular Papers)

Keywords: Causal discovery, Housing prices, Geospatial data, Clustered data, Capitalization

Abstract: This study investigates the causal effects of proximity to amenities such as schools and childcare centers on local housing prices. Understanding this relationship is critical for urban planning and policy decisions. We apply causal discovery algorithms to both synthetic and real-world housing datasets to identify potential causal relationships. These types of data provide a set of challenging properties for applying such algorithms. In particular, this study emphasizes the importance of using location as an effect modifier when dealing with clustered geospatial data. Our analysis reveals that proximity to educational facilities has a significant impact on housing prices, with variations across different postcode areas, however, this impact is not always positive. In fact the identified causal structure, indicates a negative impact of proximity of many amenities on the house prices. We conclude that causal discovery algorithms have the potential to provide novel insights into the determinants of house prices.

Serve As Reviewer: ~Arjen_Hommersom1

Submission Number: 62