Digitization of urban bicycling data

Urban bicycling data collection with mobile digital devices has become ubiquitous in cities across the globe in the last decade. However, planners, developers, data and city scientists often struggle to process, avoid biases, clean datasets, and select relevant data points, when processing, analyzing and modelling vast troves of data, when engineering applications for urban mobility. This paper aims to shed light on these data sources and bases using a descriptive and comparative analysis of the social layers embodied into this data. These ultimately and inadvertently might result in algorithmic analysis and predictive simulations with inadequate sensitivity and specificity outputs for desired requirements. Actors and social groups have vested and specific interests in these bicycling databases, which in turn shape data bases biases or disclosure policies. Therefore, their data will reflect mostly the way they interact with bicycles, cyclists and cities. How pervasive, available for academic research, and which uses are currently made with bicycling bulk data collections in the case of Berlin? Sixteen social groups that operate or hold digital bicycling data systems in that city are investigated. Results show that a clear distinction between the data produced by public and private social groups, as different uses and disclosure polices for it are identified. These contrasts significantly reduce the range of models and applications that can derive from this data. In the development of diagnostic and prognostic systems, public owned are more suitable, whereas private is for commercial application. Digitization of urban bicycling data appeared to produce different data quality and quantity. The brief description and grouping of a specific sociotechnical ensemble dynamic, such as the one conducted here, is particularly useful for the early stages of research, data collection planning and modelling, and in the design of requirements for technology development.