A hybrid data fusion methodology for household travel surveys to reduce proxy biases and under-representation of specific sub-group of population

Urban travel demand analysis efforts predominantly use household travel surveys for data supports, especially in North America. However, proxy-biases and under-representations are two dominant issues looming over the practice of using household travel surveys. Proxy biases arise from the fact that one household member reports travel diaries on behalf of all household members, often with incomplete knowledge. Under-representation arises from various reasons, including insufficient coverage of sample frame of all population segments/strata. This paper proposes a hybrid approach of data fusion between a core household travel survey dataset and data from a specialized travel survey (termed as a satellite survey) to reduce both proxy-biases and under-representation in the core dataset. Taking the Greater Toronto Area as the study area, it uses a household travel survey, the Transportation Tomorrow Survey dataset to fuse with a travel diary survey of post-secondary students in Toronto. The proposed methodology uses simulation of travel diaries (based on a dynamic econometric model of activity-travel scheduling) to correct the proxy biases of the reported travel diaries of the post-secondary students living with families and direct insertion of travel diaries (from the satellite survey) of post-secondary students living on-campus who were not well-represented in the core survey. Results prove that the proposed methodology corrects the average trip rates of this particular sub-group of the population (post-secondary students) while preserving the key travel behaviour in the core dataset.