DATA SYSTEMS
Research Group
The Data Systems Research Group (DS)
Mission
Handling large amounts of data is a challenge in almost all sectors of research and industry. Our mission is to develop innovative data systems and database technologies that are able to handle different kinds of data, such as spatio-temporal trajectory data, with scalable query processing, expressive modeling, and unified interfaces for real-world applications.
Core Research Areas
- Trajectory data models and unified spatio-temporal predicate logic with strictness levels for spatial and temporal relations under uncertainty
- Specialized operators like spatio-temporal selection, crop, and similarity joins, plus federated query optimization via predicate pushdown
- History-oblivious route and region recovery on road networks from sparse trip data, without historical trajectories
- Extensible query engines and brokers for heterogeneous data sources, leveraging modern execution frameworks
Modern data systems face challenges from different formats, inherent uncertainty, and the need to query across multiple sources, which limits analysis in areas like mobility, ecology, and logistics. Trajectory data adds to these challenges with varying storage setups, gaps between sampled points that create movement uncertainty, and a lack of common tools for space-time queries.
In the DS group, we tackle these with strong trajectory models and uncertainty-handling query methods, plus prototype systems that connect diverse sources like spatial and time-series databases. Our research combines theoretical modeling with practical query engines, delivering extensible tools for high-performance, interactive analysis of trajectories and broader data workloads.
Key Domains Include
- Urban mobility and transportation analytics on road networks
- Movement ecology and collective behavior tracking
- Logistics, environmental monitoring, and general spatio-temporal systems
Selected Prior Work and Example Projects
-
Outcome prediction and individualized treatment effect estimation in patients with large vessel occlusion stroke
SWITCH (2025) -
Interpretable Neural Causal Models with TRAM-DAGs
CLEAR (2025) -
Large language models to process, analyze, and synthesize biomedical texts: a scoping review
Discover Artificial Intelligence (2024)
-
Bernstein flows for flexible posteriors in variational Bayes
AStA (2024) -
Deep learning versus neurologists: functional outcome prediction in LVO stroke patients undergoing mechanical thrombectomy
Stroke (2023) -
Short-term density forecasting of low-voltage load using Bernstein-polynomial normalizing flows
IEEE Transactions on Smart Grid (2023)
Leadership
Members
Publications
Loading publications...