Deep learning geodemographics with autoencoders and geographic convolution

We present two approaches to creating geodemographic classifications using deep neural networks. Both deep neural networks are based on autoencoders, which allow automating dimensionality reduction before clustering. The second approach also introduces the idea of geographic convolution in neural networks, which aims to mirror in the geographical domain the approach of graphical convolution, which has revolutionised image processes in the past decade. To test our approaches, we created a geodemographic classification based on the United Kingdom Census 2011 for the county of Leicestershire and compared it to the official 2011 Output Area Classification. Our results show that the two deep neural networks are successful in creating classifications which are statistically similar to the official classification and demonstrate high cluster homogeneity.