Projects

Active projects

  • NOW: Synopses for data streams — We design novel synopses for summarization of fast-paced data streams. Recent examples of our work include OmniSketch (VLDB’24, VLDBJ’26) and SpatialSketch (EDBT’25). Most of our code is open-source. The project is partially funded by EU Horizon Europe programme (STELAR, No. 101070122), NextGenerationEU (ARCHIMEDES Unit, OPS 5154714), and TU/e.
    Key international collaborators: Minos Garofalakis.
  • Multivariate Similarity search — Identifying similarities between data objects in a dataset is a fundamental task in data science, with applications across a wide range of disciplines. The group develops a new class of algorithms for similarity search, focusing on (i) high-order similarities, which involve relationships among three or more time series, and, (ii) pairwise similarity measures for multivariate (multi-channel) time series. Examples of our work include MSIndex (VLDB’26) and Correlation Detective (VLDBJ’23, VLDB’22). The project is partially funded by EU Horizon Europe programme (STELAR, No. 101070122).
    Key international collaborators: Themis Palpanas, John Paparrizos.

Past projects

  • STELAR: Spatio-TEmporal Linked data tools for the AgRi-food data space — STELAR developed a knowledge lake management system fit for today’s agrifood space. The project received funding from the EU Horizon Europe programme.
  • SmartDataLake: Sustainable Data Lakes for Extreme-Scale Analytics — SmartDataLake designed, developed and evaluated novel approaches and techniques for extreme-scale analytics over Big Data Lakes, facilitating the journey from raw data to actionable insights. The project was funded by EU H2020 programme.

Software

  • SynLib — We are maintaining a repository of open-source implementations of synopses (mostly in Java) for analyzing streaming data. Get it here. Contact us if you are interested to hear more.
  • OmniSketch — OmniSketch synopsis supports analytics over fast-paced data streams, similar in nature to the traditional OLAP analytics in relational databases – i.e., with aggregates and selection predicates. The last version also provides support for streams with deletes (the turnstile model).
  • Correlation Detective — Correlation Detective is the only scalable solution for identifying high-order correlations in large datasets. You can use it online for small datasets, or download it and use it locally for big datasets. Contact us if you would like to see a guided demo on your own data.