Another goal of the study was to identify typical user groups (households) in the regional samples sharing specific characteristics pertaining to the modal choice and its most relevant factors. As a multivariate method for multidimensional data, statistical two-step cluster analysis with SPSS was used to identify groups and relevant criteria of modal choice segmentation in the transport systems and respective demographic, socio-economic and spatial contexts. Two-step cluster analysis also automatically selects the number of homogeneous clusters. The two-step clustering method was applied separately to the data sets containing only valid, complete and plausible cases (responses) from Austria (N=240), Finnish Growth Corridor (N=456), and Basel Region (N=166). The main research questions addressed with cluster analysis were:
- Are there, and what kinds of, distinguishable user (commuter) groups?
- What are the most prominent determinants of modal choice in commuting? For example, to what extent does the disposable income influence commuting behaviour?
- How do the interdependencies and dynamics compare across the three case areas in Austria, Finland, and Switzerland?
Variables for the cluster analysis were household size, residential location, the number of children, dominant modal choice and workplace location. Household income and single parent family as a household structure turned out to be less of importance as a predictor.
Based on these seven variables, the clustering algorithm produced five internally homogenous clusters of approximately same size both in Austria and Switzerland, while Finland had seven different clusters. As an example, the variables and their relative importance for cluster formation for Austria are illustrated in the figure below.

Profile 1 consists of urban single person households predominantly using public transport and to a lesser extent active modes (walking and cycling). Their workplace is located in the urban area, and they have relatively low total household income.
Profile 2 consists of couples or small family households (2-3 persons) living in a semi-urban environment and working in the city centre. They are typically car-reliant, but occasionally engaging in active modes. They share a relatively low household total income.
Profile 3 consists of large-family households located in a rural environment and heavily relying on a car. They have urban workplaces and average total income.
Profile 4 consists of well-off urban couples or small family households. They work and reside in urban areas or city centres and have high total income. They are a) heavily reliant on a car and b) the least active (walking/cycling) compared to other profiles.
Profile 5 consists of urban couples or small family households with a very high affinity towards public transport and active modes. They have an average income.
In general, the Swiss clusters of commuting profiles are highly similar, the main differences being in the commuting modes – higher car reliance in Austria and higher shares of active modes in Basel Region. As this difference owes largely to the spatial differences of these regions, it can be concluded that these regions share some cultural characteristics, which may be linked to their geographic proximity.
The composition of user groups in Finland differs from Austria and Switzerland by a greater degree of heterogeneity within the sample, which explains the higher number of clusters – seven – but also by a different factor importance for cluster formulation in the clustering process. The dominant modal choice plays much higher importance, which means that it is more a cause than just a consequence, which is a clear contrast with the other regions. The fact that a private car is used by singles and small families in commuting to urban working locations indicate that the choice of a private car is justified by other than rational grounds only. In other words, the data supports the conclusion that more than in the other surveyed countries, a private car is associated with specific cultural values in Finland, e.g., private car as a status symbol. While the user profiles in Finland shows some distinctiveness in comparison to the other regions, the homogeneity within the user groups in Finland highlight the explanatory power of the chosen background variables.
Based on the cluster analysis, the existence of relative homogenous user groups among commuters suggests that specific needs concerning transport exist within each group. Accordingly, while the enablers for increasing the use of public transport or emerging services consist of lower ticket prices, more frequent services, reduced waiting times, etc. at a general level, they will have substantially different impacts on the user groups. This means that the likelihood of modal shift subject to the introduction of new services will be different for each group: while some groups may be quick to take up new commuting services, others may totally ignore them.
It is logical to assume that also the future modal choices are influenced by the cluster factors used in our study. Knowing about the cause and effect patterns behind modal choice will significantly support the development of new services and help improve their market potential. When designing new services to commuters, transport service providers are well advised to ‘know’ their potential customers well regarding the socioeconomic and demographic characteristics as well as the operational area of the services. On this basis, new services can be customized to fit the respective commuters’ needs, and market potentials can be assessed more realistically.