Visual Interactive Space-Time Segmentation

There are numerous applications where data is recorded at specific geographic locations over certain periods and analyzed by experts. Examples include commerce (e.g., sales of items in shops), meteorology (e.g., climate and weather data), demographics (e.g., household wealth), or epidemiology (e.g., Covid cases). To identify correlations and underlying patterns, data are segmented into coherent spatial, temporal, or spatiotemporal areas. This allows for targeted investigation of large and complex datasets and helps to derive insights and knowledge from them. However, spatial and temporal variations may lead to data quality issues, such as uncertainties or random errors, which can only be resolved with human experience and background knowledge. In our project, we integrate human background knowledge into the segmentation and data exploration processes with spatiotemporal data, so that they may work correctly despite existing errors and uncertainties. This is achieved through novel combinations of expert knowledge, visualizations, interactions, and algorithms. The resulting segmentations allow for unexpected and new insights into the data and underlying processes, even in the presence of data quality issues. The goal of our project is to achieve significant progress in the segmentation of spatiotemporal data, especially when dealing with erroneous data. Additionally, we aim to enhance the knowledge generation process through the seamless integration of expert knowledge into visualizations and algorithms. Furthermore, we will demonstrate how these insights can be applied to other application areas.