Cross-analysis of the variability of travel behaviors using one-day trip diaries and longitudinal data

Many travel surveys used in transportation planning still collect information over a single weekday while mobility and activity patterns fluctuate from day to day. However, the emergence of new continuous data sources has led to the generation of large amounts of longitudinal data. Therefore, this paper aims to value these two data types to cross-analyze the temporal variability of aggregated travel behaviors. A weighting method based on large-scale single-day survey samples was first developed to extend the representativeness of the daily trip diaries available in the three most recent Origin-Destination household travel surveys of Montreal, Canada, to the entire data collection period (four months in the fall). This method was then validated using passive data streams (smart card and count data) for three modes: car, subway, and bicycle. The variations in the daily use of these modes measured independently by the surveys and passive data were compared over several months of the fall period and throughout an average week. Indicators were also proposed to evaluate whether the two data sources capture the same trend and seasonality, especially the weekly rhythms. The results show that it is possible to accurately infer day-to-day variability in travel behavior from a large-scale cross-sectional single-day survey. Furthermore, this paper demonstrates the potential for complementarity between traditional surveys and emerging data (at least over the survey period), as well as the possibility to combine them since there are large similarities in the aggregated travel patterns they allow monitoring.

Automated summary

This paper evaluates the value of traditional surveys and emerging data in cross-analyzing the temporal variability of travel behaviors. The results show that day-to-day variability in travel behavior can be accurately inferred from a single-day survey, and demonstrate the complementarity and potential combination of the two data sources.