Uncertainty and Trust-Aware Integration of VGI and Spatio-Temporal Traces for Understanding Animal Behavior

Recent advances in biologging technology enable the wholesale, accurate, and reliable recording of animal movements on a global scale. Though, such data often do not entail sufficient information on semantic context and quantification.

In this project, we address the afore issue by matching Movebank tracking data with volunteered geographic information (VGI, here:  human observations) from different web portals (e.g., eBird, GBIF). We implement semi-automated methods to assess and visualize the uncertainties of both VGI data and the actual matching process. Thus movement ecologists will visit a visual-interactive web platform, enabling them to visually explore the matching result that is followed by iterative and uncertainty-aware refinements. Subsequently, movement ecologists will be able to enrich existing behavior prediction models by suitable VGI data, while following once again a visual-interactive feedback loop.

Research Questions

  • Which semi-automated matching strategies are most relevant for matching animal
    movement trajectories with VGI?
  • Given VGI and the matching process, how can uncertainties be evaluated and visually
    communicated in the most effective ways?



This research project receives funding from the DFG (Deutsche Forschungsgemeinschaft) within the priority program “Volunteered Geographic Information: Interpretation, Visualisation and Social Computing” (SPP 1894).