Enhancing human mobility research with open and standardized datasets

The proliferation of large-scale, passively collected location data from mobile devices has enabled researchers to gain valuable insights into various societal phenomena¹. In particular, research into the science of human mobility has become increasingly critical thanks to its interdisciplinary effects in various fields, including urban planning, transportation engineering, public health, disaster management, and economic analysis². Researchers in the computational social science, complex systems, and behavioral science communities have used such granular mobility data to uncover universal laws and theories governing individual and collective human behavior³. Moreover, computer science researchers have focused on developing computational and machine learning models capable of predicting complex behavior patterns in urban environments. Prominent papers include pattern-based and deep learning approaches to next-location prediction and physics-inspired approaches to flow prediction and generation⁴.

Regardless of the research problem of interest, human mobility datasets often come with substantial limitations. Existing publicly available datasets are often small, limited to specific transport modes, or geographically restricted, owing to the lack of open-source and large-scale human mobility datasets caused by privacy concerns⁵. Examples of real-world trajectory datasets include the widely used GeoLife⁶, T-Drive trajectory dataset⁷, the NYC Taxi and Limousine Commission dataset⁸, and the Gowalla dataset⁹, and although such datasets are valuable in conducting large-scale experiments on human mobility prediction, the lack of metropolitan-scale and longitudinal open-source datasets of individuals has been one of the key barriers hindering the progress of human mobility model development. The lack of open data also perpetuates gatekeeping, where researchers without access to exclusive datasets are excluded from this research area, raising equity concerns in science. Moreover, even in the case where researchers may access processed mobility datasets, privacy concerns limit access to raw and open data sources. This means that even the datasets that are publicly available are often pre-processed without using standardized procedures. It is possible to obtain a completely different dataset just by slightly changing a parameter in the data pre-processing pipeline, for instance, by changing the spatial and temporal definition of a stop location. This makes it difficult to conduct fair performance comparisons across different methods¹⁰.