Understanding the effects of socioeconomic factors on housing price appreciation using explainable AI

Abstract

Housing price appreciation is an important socioeconomic phenomenon that captures the complex socioeconomic dynamics of a city. Variation in housing price appreciation across neighborhoods reflects localized housing demand and supply-side factors. This study develops quality-adjusted, census tract-level housing price indices using a fine-grained big dataset containing a total of 140,289 housing transactions in the County of Los Angeles. We employ the SHapley Additive exPlanations (SHAP) technique, an explainable artificial intelligence framework, to examine the underlying demographic and socioeconomic factors that help in explaining the variance in tract-level housing price appreciation from 2012 through 2018 in the County of Los Angeles. The novelty of the methodology lies in the local interpretation of spatial patterns it provides from big data in the urban context and in assessing how the factors influencing housing price appreciation vary geographically. The modeling framework could help planners in making informed decisions about local geographic contexts that contribute to variability in housing price appreciation in cities.