Applying machine learning methods to housing listings data and satellite imagery to measure the spatial and temporal structure of housing sub-markets

My PhD uses new forms of data to understand the spatial structure and differences between housing sub-markets (areas of similar housing prices) in the capital city of Spain. The PhD is so far composed of two research papers (listed below). The final paper will explore variations in the drivers in housing prices using the sub-markets identified from the second paper. The PhD is partnered with real estate listings portal Idealista, who provided some of the data for the thesis.

Intra-Urban House Prices In Madrid (Spain) Following The Financial Crisis: An Exploration Of Spatial Inequality
In this paper we explore the temporal dynamics of spatial inequality in housing prices for Madrid, the capital city of Spain. Spatial inequalities are a concerning feature of urban areas across the globe. It has been suggested that within cities, housing prices are becoming more geographically unequal over time, particularly since the 2008 housing market crash. However, more evidence is needed at the intra-urban level to understand neighbourhood house price differences in large urban areas. Changes are analysed during a key period of housing market bust (2010-2015) and boom (2016-2019), using data from a major housing listing portal. Fine grain space-time analysis of the distribution of housing prices supports an increase in spatial inequality and polarisation at the neighbourhood level. Two spatially differentiated housing sub-markets of high- and low-priced housing are identified. The persistence and growth of spatial house price inequality has important societal implications for the wealth gap and segregation of rich and poor in cities.

Extracting Features from Satellite Imagery to Understand the Size and Scale of Housing Sub-Markets in Madrid
The following paper proposes a novel machine learning approach to the segmentation of urban housing markets. We extract features from globally available satellite imagery using an unsupervised machine learning model called MOSAIKS, and apply a k-means clustering algorithm to the extracted features to identify sub-markets at multiple intra-urban scales within a case study of Madrid (Spain). To systematically explore scale effects on the resulting clusters, the analysis is repeated with varying sizes of satellite image patches. We assess the resulting clusters across scales using several internal cluster-evaluation metrics. Additionally, we use data from online listings portal Idealista to measure the homogeneity of housing prices within the clusters, to understand how well sub-markets can be differentiated by the image features. This paper evaluates the strengths and weakness of the method to identify urban housing sub-markets, a task which is important for planners and policy makers and is often limited by a lack of data. We conclude that the approach seems useful to divide large urban housing markets according to different attributes and scales.