• Travel behaviour is critical

We are happy to announce the public release of the open dataset “MobilityNet”, a collection of data related to transparent and privacy aware mobility tracking through an open-source app. The dataset was accepted and presented as a poster at the Workshop “Tackling Climate Change with Machine Learning”, in the frame of the International Conference on […]

We are happy to announce the public release of the open dataset “MobilityNet”, a collection of data related to transparent and privacy aware mobility tracking through an open-source app. The dataset was accepted and presented as a poster at the Workshop “Tackling Climate Change with Machine Learning”, in the frame of the International Conference on Learning Representations (ICLR), one of the most well-appreciated conferences on machine learning and artificial intelligence. Since the entire conference took place virtually, you can still drop by and watch our 12-minute poster presentation at the workshop whereas our description paper is freely available in PDF form.

The publication of the dataset was motivated by the focus on reducing CO2 emissions related to mobility, which can be controlled by influencing transportation demand. Individual user mobility models are key to influencing demand at the personal and structural levels. Constructing such models is a challenging task that depends on a number of interdependent steps, and unfortunately progress on this task is hamstrung by the lack of high quality public datasets.

As a response to that need, MobilityNet is the first step towards a common ground for multi-modal mobility research. MobilityNet solves common problems faced by similar mobility dataset collections, such as the holistic evaluation, the privacy preservation and the fine-grained ground truth. In order to achieve that, it makes use of artificial trips, control phones, and repeated travel, respectively.

The current dataset includes 1080 hours of data from both Android and iOS, representing 16 different travel contexts and 4 different sensing configurations. We hope that this dataset can be used to facilitate further research in the field. The release was led by K. Shankari (then UC Berkeley, now NREL) main developer of the open source mobility app “E-mission”, having as co-authors Jonathan Fürst and Mauricio Fadel Argerich (NEC Laboratories Europe), Jesse Zhang (UC Berkeley) and Eleftherios Avramidis from the Open Source Lab for Sustainable Mobility (DFKI).