Law, S; Chen, PN; Shen, Y; Karimi, K; Penn, A; (2022) Estimating house price with spatial and land use accessibility components using a data science approach at the national scale. In: Proceedings 13th International Space Syntax Symposium, SSS 2022. Western Norway University of Applied Sciences (HVL)

Preview

PDF 406law.pdf - Published Version Download (6MB) | Preview

Abstract

Extensive research had been conducted studying the spatial configuration effects on house price using the hedonic price approach. Previous research has mostly focused on using econometric approaches in estimating house price. With the growing popularity of machine learning methods, there is an opportunity to study this problem from a data science perspective. Following Law et al (2017) which studied how economic value of closeness centrality (integration) differed across cities in England, we conduct here a similar experiment examining these differences using a data science approach. We leveraged on an integrated urban model, a large-scale geographic database to compute a series of land use accessibility and space syntax accessibility measures at the country scale (~120 measures). We then use a compressed set of spatial and land use accessibility components to estimate a set of hedonic price models in England; i. first for the entire country, then ii. for all 22 cities and then iii. for 22 cities individually. We found that spatial and land use accessibility features improve house price prediction accuracy jointly and the improvements are greater when using nonlinear methods. This research serves as a basis on the application of data science approaches in space syntax research for predicting real estate outcomes at the National-Scale.

Download activity - last month

Export as XMLCSVJSON

Download activity - last 12 months

Export as XMLCSVJSON

Downloads by country - last 12 months

Export as XMLCSVJSON

Archive Staff Only

View Item