Abstract

In this era of data, extensive information at urban-scale is collected yet thoroughly analysed. To assist the analysis of urban data in a more integrated way, we proposed a two-layer cluster analysis framework for policymakers by taking advantage of the aligned design of geographical boundaries. The framework performs two layers of cluster analysis with algorithms selected according to the scale and purposed of each desired resolution, and variables selected via both correlation and predictive power. A case study of domestic gas consumption patterns in the Greater London Area is conducted to illustrate the framework. We generated clusters at both small-area and local authority level based on the data of social economics, energy consumption, built environment, etc. With no gas consumption and geographical information directly included in the modelling process, each small-area cluster shows distinct domestic gas consumption and geographical characters. Together with the differences among local authority clusters, we provided a new perspective for policymakers and various gas consumption and energy efficiency policy implications.