Enriched travel demand estimation by including zonal and traveler characteristics and their relationships

Current procedure in travel demand estimation models is to separately deal with attraction, production and trip distribution, where the latter typically assumes inverse distance proportionality. We show that this procedure leads to errors in the demand estimation, particularly when dealing with very specific zones and heterogeneous travel behavior. We argue that this traditional procedure is rooted in traditional ways of data collection, while new (big) data sources allow direct observation of travel demand patterns. Using such data, we propose an enriched travel demand estimation method in which zonal and traveler characteristics and their relationships are consistently carried over from the empirical data into the demand model. This can improve both the validity and richness of demand estimations.