Defining Street-based Local Area and measuring its effect on house price using a hedonic price approach: The case study of Metropolitan London

- The definition of local area unit is an under-explored topic in housing studies.
- Applying community detection on the dual graph to defining Street-based local area.
- Research links street network community structure with house price.
- Results showed significant local area effects on house prices.
- Results showed Street-based Local Area is preferred to region-based one.

An under-explored topic within the field of planning and housing studies is related to the definition of local area unit. An empirical problem that arises is that different types of local area units can infer different results. This could be in constructing segregation indices, in estimating hedonic price models or in identifying housing submarkets. This research proposes the concept of Street-based Local Area (SLA), in asking to what extent SLA associate with house price. In order to examine this question, this article borrows from network science and space syntax research in defining SLA. This research conjectures that SLA has a significant effect on house price and that this effect is captured more strongly than ad-hoc administrative region-based local area. In order to test this conjecture, this research adopted the multi-level hedonic price approach to estimate local area effects on house prices for the case study of Metropolitan London in the United Kingdom. Results showed significant local area effects on house prices and that SLA is preferred to region-based one. The plausible reasons are firstly, people perceived the local area on a street network. Street-based Local Area is able to capture more precisely subtle perceptual differences in an urban environment than an ad-hoc administrative region. Second, the topology of the street network reinforces the socio-economic similarity/differences overtime. Differences between local areas can become more pronounced as like-minded people bump into each other, cluster together and share information with each other. Third, as people identify these local areas they would make decisions based on it. The local area becomes part of the housing bundle leading to it having an effect on house price. The main contribution of the research is the novel application of community detection techniques on the street-network dual graph to defining SLA. This is important as it links the topology of the street network to how we define and perceive local area and it presents an alternative to ad-hoc administrative geographies that are currently applied in many aspects of neighbourhood planning.