Geodemographics, visualisation, and social networks in applied geography
Alexander D. Singleton; Paul A. Longley (2009). Applied Geography, 29(3), 289-298. DOI: 10.1016/j.apgeog.2008.10.006
Abstract
This review begins by acknowledging the success of geodemographics as an important area of activity in applied geography. However, it then develops a critique of the conceptual and computational underpinnings of the approach, and argues that changes in data supply and online communication have rendered current practices obsolete. It presents elements of a new perspective, entailing: changes in the specification, estimation and testing of online geodemographic systems; adoption of consultative practices from online folksonomies; automated generation of pen portraits; and ‘on the fly’ visualisation of the outcome of geodemographic classifications.
Extended Summary
This research examines how traditional geodemographic classification systems - which categorise neighbourhoods by social, economic and demographic characteristics - need fundamental transformation to remain relevant in the digital age. The paper traces geodemographics from their origins with the Chicago School in the 1920s through to modern commercial applications in marketing and public service delivery, noting their apparent success over 30 years of use. However, the research argues that current proprietary systems have not evolved due to inherent superiority but rather represent outdated responses to past limitations in data supply, computing power, and disclosure controls. The study critiques existing geodemographic approaches on several fronts. Current systems fail to incorporate spatial considerations adequately, treating geography merely as units for classification rather than recognising local contextual effects. The methodology reveals that most commercial classifications operate as ‘black boxes’ where end users cannot scrutinise the data sources, variable weightings, or clustering methods used. This lack of transparency undermines scientific reproducibility and prevents proper validation of results. Furthermore, the research highlights how technological changes have transformed the landscape for creating geodemographic systems. The advent of free public sector data, open source statistical software, and web-based data sharing protocols has democratised access to the tools needed for classification development. Web 2.0 technologies now enable real-time data updates and interactive visualisation through platforms like Google Maps APIs. The paper proposes a revolutionary new model centred on application-specific, consultative approaches rather than general-purpose classifications. This vision includes online interfaces where users can build bespoke classifications using palette-based tools, incorporating variable weighting through slider controls and real-time preview capabilities. The research advocates for classifications that are engineered for particular applications, scientifically reproducible, and incorporate stakeholder participation through consultation exercises similar to online folksonomies. The study demonstrates these concepts through examples like the Public Profiler website, which uses external APIs to visualise typologies and enable public feedback on classification accuracy. This approach moves beyond post-facto rationalisation toward genuine community engagement in the classification process. The broader significance extends to both academic geography and practical applications in marketing and public policy. For commercial applications, the research suggests moving away from ‘one size fits all’ classifications toward problem-centred approaches where different classifications serve specific purposes - recognising that systems effective for predicting holiday preferences may be inappropriate for understanding attitudes toward public health services. In the public sector, this transformation could enable more sophisticated area classification methods that unlock the value of locally collected data, moving beyond crude measures toward nuanced understanding of neighbourhood dynamics and service needs.
Key Findings
- Current geodemographic systems operate as ‘black boxes’ lacking transparency in methodology, data sources and variable weighting schemes.
- Technological advances including free public data, open source software and Web 2.0 protocols have democratised geodemographic classification capabilities.
- Application-specific classifications outperform general-purpose systems for predicting domain-specific behaviours like public service attitudes versus consumer preferences.
- Stakeholder consultation and folksonomy approaches can improve classification accuracy and public acceptance through community engagement.
- Real-time data integration and interactive visualisation tools enable ‘on the fly’ classification generation tailored to specific user requirements.
Citation
@article{singleton2009geodemographics,
author = {Alexander D. Singleton; Paul A. Longley},
title = {Geodemographics, visualisation, and social networks in applied geography},
journal = {Applied Geography},
year = {2009},
volume = {29(3)},
pages = {289-298},
doi = {10.1016/j.apgeog.2008.10.006}
}