Residential neighbourhood classification: An environmentally enhanced approach

National small area classifications in Britain were first produced over 40 years ago using statistics from 1971 Population Census and have now become a regular feature of governmental, academic and commercial analysis of census information. These classifications aim to encapsulate the aggregate demographic and socio-economic character of small areas by means of a simple thumbnail description. However, these pen portraits often also refer to the environmental nature of the different types of area where people live, employing terms such as ‘leafy suburb’, ‘industrial hinterland’ or ‘agricultural heartland’. This paper reports on research that aims to determine whether a set of environmental (land use) indicators are capable of discriminating between areas in a way that matches a ‘standard’ area classification derived from multivariate analysis of demographic and socio-economic statistics. The research assesses the impact of adding a set of environmental (land use) variables to a collection of Census variables on area classification using k-means clustering in two contrasting case study local authorities. The results reveal that clustering with and without the addition of land use variables produce partially overlapping (coincident) classifications of the small areas and certain of the land use variables are aligned with some area types.