Combining a regional household survey and passive data streams for longitudinal monitoring purposes

More and more passive big data are available, but travel surveys are still widely used in transportation. This coexistence brings questions about the role of each data source and new data fusion challenges. Within this context, this paper aims to combine the Montreal household survey and passive data streams to provide longitudinal monitoring of aggregated travel behaviors. The proposed methodology, based on time series decomposition, consists in applying the trend and seasonal variations detected in passive data to the typical fall weekday observed in the survey. This allows traditional mobility indicators, such as modal shares, to be projected and annualized.