Traveler segmentation strategy with nominal variables through correspondence analysis

Travel research is increasingly exploring the role of qualitative elements in mobility behaviors, beyond customarily considered factors such as the socioeconomic status of travelers or the characteristic and performances of transport systems. More recent travel surveys are thus increasingly collecting non-metric information through categorical variables such as opinions, which is difficult to exploit with the most widely used analytical tools. The present paper assesses the potential of a nonparametric analysis technique little used in transport research, namely correspondence analysis, in order to define a set of different customer profiles regarding stated mode choices. Data coming from an attitudinal travel survey administered to a representative sample of the population of the city of Novara (Italy) are used to this effect. The resulting clusters are quite informative and policy relevant, mainly because they are based on a variety of metric and nominal variables which would be less easy to consider when using more standard multivariate techniques such as cluster analysis, and also because not all the observation need to be forcedly classified. A simple modal choice modeling exercise illustrates how the derived market segments can provide guidance to improve the results of a standard quantitative analysis, while keeping a low computational complexity. Our study shows the usefulness of the proposed methodology in a transport policy decision-making context.