Modeling Human Spatial Behavior Through Big Mobility Data

People are engaged in a variety of activities through space every day. The choice of type and location of activities is known as human spatial behavior. Urban decision makers need to understand how land use and transportation systems can shape human spatial behavior in order to design better systems. In the past, they have collected mobility data through travel surveys to understand human spatial behavior. Today, a wide range of automatically collected data have become available as alternative data sources.
Big mobility data vs. traditional travel survey data has been a topic of long-time debate in human mobility and travel behavior research. Big data are intuitively better but this is not always the case. Big mobility data relate to a large number of travelers and trips but little is known about each individual individual traveler and trip, not to mention that sometimes their information has to be aggregated for privacy concerns. On the other hand, travel survey data, despite reporting only a small group of respondents, tend to include abundant features about each individual traveler, such as age and attitudes, and each trip, such as trip purpose. Assuming that each row represents one traveler and each column represents one feature, big mobility data should have been described as long and thin, and “small” survey data (Chen et al., 2016) as short and wide.