Understanding public transit patterns with open geodemographics to facilitate public transport planning

Abstract

Plentiful studies have discussed the potential applications of contactless
\nsmart card from understanding interchange patterns to transit
\nnetwork analysis and user classifications. However, the incomplete
\nand anonymous nature of the smart card data inherently limit the
\ninterpretations and understanding of thefindings, whichfurther limit
\nplanning implementations. Geodemographics, as ‘an analysis of people
\nby where they live’, can be utilised as a promising supplement
\nto provide contextual information to transport planning. This paper
\ndevelops a methodological framework that conjointly integrates personalised
\nsmart card data with open geodemographics so as to pursue
\na better understanding of the traveller’s behaviours. It adopts
\na text mining technology, latent Dirichlet allocation modelling, to
\nextract the transit patterns from the personalised smart card data
\nand then use the open geodemographics derived from census data
\nto enhance the interpretation of the patterns. Moreover, it presents
\nnight tube as an example to illustrate its potential usefulness in
\npublic transport planning.