Scalable Online Data Management (SODATA) group, TU/e

The SODATA group in the Department of Mathematics and Computer Science at Eindhoven University of Technology aims to enable interactive and scalable data science in an era of ever-growing data volumes and increasingly fast data streams. Our vision is to make large-scale data science responsive by design, allowing users and systems to reason over massive, continuously evolving data in near real time, typically with millisecond- to sub-second latency.

We focus on the design of algorithms and data structures that deliver responses within (milli)seconds, enabling meaningful human interaction with data and timely automated decision making. Such responsiveness is essential to keep users engaged in the data science workflow, to ensure results remain current in fast-paced streaming environments, and to support immediate action in time-critical domains such as security monitoring and fraud detection.

To realize this vision, we develop approximate algorithms and compact data structures – typically called synopses – that summarize high-volume data streams and time series using limited computational resources. In parallel, we design and implement distributed algorithms that scale efficiently across hundreds of machines. Our research forms the foundation for applications ranging from sensor network analytics and large-scale similarity detection to, more recently, efficient and scalable AI training.

Keywords

• Streaming algorithms and analytics
• Approximate query processing with synopses
• Time-series data management
• Scalable data systems

Collaboration opportunities

We actively seek collaborations with academic partners, research institutes, and not-for-profit organizations to tackle challenging problems aligned with our expertise. We also collaborate closely with industry, both through EU-funded projects and company-funded research. If you are interested in exploring a collaboration, we would be happy to hear from you.