Abstract:

To realize the full potential of crowdsourced data collected by smartphone applications in urban research and planning, there is a need for parsimonious, reliable, computationally and temporally efficient data processing routines. The literature indicates that the opportunities brought by crowdsourced data in generating low-cost, bottom-up, and fine spatial and temporal scale data, are also accompanied by issues related to data quality, bias, privacy concerns and low accessibility. Using an exemplar case of RiderLog, a crowdsourced GPS tracked bicycling data, this paper describes and critiques the processes developed to transform this urban big data. Furthermore, the paper outlines the important tasks of formatting, cleaning, validating, anonymizing and publishing this data for the capital cities of each state and territory in Australia. More broadly, this research contributes to the foundational underpinnings of how to process and make available crowdsourced data for research and real world urban planning purposes.

Keywords: