Activity travel of population segments grouped by daily time-use: GPS tracking in Halifax, Canada

This paper examines the weekday activity patterns and associated travel behaviour of urban residents grouped by similarity of their daily time-use profiles. We analyze the two-day in-home and out-of-home activity of respondents to the large STAR time-use and travel survey. The efficient Fuzzy C-Means (FCM) clustering algorithm yields 12 clusters of activity days grouped by homogeneous activity patterns. Six clusters are characterized as out-of-home workers, five as non-workers, and one as students. Time-stamped GPS-tracking of activity is employed to map out-of-home locations of individuals at key times in the day, which reveals considerable variation between the groups in their locational and temporal preferences. Worker groups cluster in employment areas, with 9-5 workers focused downtown and earlier-shift workers in suburban business parks. Non-workers are more dispersed, and more clustered in retail/service areas. Time-diary data also show clear variations in typical triptour behaviour, relating both to trip purposes and to tour complexity: non-worker groups take more tours, on average, but fewer complex tours.