SLH: Understanding Travel Demand through Passively-generated Mobile Data: a Python-based Mobility Analysis Workshop

The ubiquity of GPS-equipped mobile devices has enabled the collection of human mobility data with high spatiotemporal granularity. Indeed, there now exists an ecosystem of both data providers and consulting agencies centered around collecting, processing, and extracting insights from this type of data. Though much of the raw data cannot be made public due to privacy-preserving agreements, the academic community can still access geographically aggregated data. This workshop will cover the (Python-based) preprocessing and wrangling of aggregated mobility data in New York state provided by SafeGraph. It will highlight concepts including (but not limited to) origin-destination matrices, the four-step travel demand model, self-selection bias in passively-generated data, and demand prediction. It will also describe how to fuse such data with the American Community Survey (ACS) providing sociodemographic information at the census block group-level. Relevant packages that will be leveraged and explained include GeoPandas and scikit-Mobility. Intermediate familiarity with Python, data structures, and object-oriented programming is recommended.