Housing submarkets can be defined as a set of dwellings that are reasonably close substitutes with one another, but poor substitutes between other submarkets. This research argues similarities within submarkets are not only captured by its building and location characteristics but also in how each dwelling is inter-connected within its local area and embedded to the rest of the system. This research conjectures that spatial network local-areas as defined by community detection methods can be used to identify spatial housing submarkets. In order to test this conjecture, the hedonic approach will be used as an empirical strategy on the case study of London. The study found spatial network local areas correspond with planned known local area boundaries and that greater house price similarity is found within spatial network local-areas than between. The study also found that spatial network local area as defined by community detection technique can be used to identify spatial housing submarkets to explain house price. The contribution of this research is it represents a proof of concept in the use of community detection techniques in the definition of spatial housing submarket. Importantly it illustrates the significance in how spatial configuration influences housing market not just in terms of accessibility (Law et al. 2013) but also in terms of housing submarket. Further research will be carried out to study the spatial configuration of the spatial network local areas in understanding severances and connectivity between them. By understanding cities through multiple spatial representations will allow more informed policies at the local-area level.