Upset diagrams for examining whether parking maximums influence modal choice and car holdings

Abstract

Venn and Euler diagrams can provide researchers and practitioners with a quick visual summary of information that is relevant for evaluating policy, but given that the potential number of intersections grows exponentially relative to sets of information (2 n – 1), diagrams featuring more than three sets of information are typically difficult to interpret from a two-dimensional medium (i.e. reports or articles). In contrast, upset diagrams arrange sets and intersections on a matrix. Thus, the potential number of intersections grows linearly, and further information (e.g. histograms and boxplots) can be aligned vertically or horizontally to this matrix. To demonstrate this application, seven commuter modal choices, two car-parking policies, car holdings and country of birth will be visualised using an upset diagram to evaluate whether a parking-maximums policy reform in Brisbane, Australia, is reducing car ownership and dependence.