ARMADA

Complex analysis over heterogeneous data sources has become the basis of critical decision making. Nowadays, for numerous analytics tasks, actual content matters more than size. Analysts are baffled by the amount of available sources in search for the "right data" that maximizes analytics performance giving their company a competitive edge. The effect that different data inputs have on analytics tasks is still a missing piece in the puzzle. The ARMADA project (Boosting Analytics Performance: A Data-driven Approach) plans to examine and export a quantitative link between data and the analytics workflows that consume them. This work studies the novel notion of utilizing dataset similarity to infer operator behavior and be able to build scalable, operator-agnostic performance models for real-life popular big-data tasks.

Data profiler is our open-source tool for modeling various analytics operators over multi-dataset inputs.

• D. Tsoumakos and I. Giannakopoulos: Content-based Analytics: Moving Beyond Data Size. Accepted for publication in the The Sixth IEEE International Conference on Big Data Computing Service and Machine Learning Applications, Apr. 13 - 16, 2020, Oxford, UK.
• T. Bakogiannis, I. Giannakopoulos, D. Tsoumakos and N. Koziris: Apollo: A Dataset Profiling and Operator Modeling System. In Proceedings of the 2019 ACM SIGMOD/PODS International Conference on Management of Data, June 30 - July 5, 2019 Amsterdam, The Netherlands.
• T. Bakogiannis, I. Giannakopoulos, D. Tsoumakos and N. Koziris: Predicting Graph Operator Output over Multiple Graphs. In Proceedings of the 19th International Conference on Web Engineering, June 11 - 14, 2019, Daejeon, Korea.
• I. Giannakopoulos, D. Tsoumakos and N. Koziris: A Content-Based Approach for Modeling Analytics Operators. In Proceedings of the 27th ACM International Conference on Information and Knowledge Management (CIKM 2018), October 22-26, 2018, Lingotto, Turin, Italy.