Demand Modelling for Responsive Transport Systems Using Digital Footprints

Traditionally, travel demand modelling focused on long-term multiple socio-economic scenarios and land-use configurations to estimate the required transport supply. However, the limited number of transportation requests in demand-responsive flexible transport systems require a higher resolution zoning. This work analyses users short-term destination choice patterns, with a careful analysis of the available data coming from various different sources, such as GPS traces and social networks. We use a Multinomial Logit Model, with a social component for utility and characteristics, both derived from Social Network Analyses. The results from the model show meaningful relationships between distance and attractiveness for all the different alternatives, with the variable distance being the most significant.