Jonas Traub is a doctoral candidate at the Database Systems and Information Management group (DIMA) at TU Berlin. In March 2017, he presented the demonstration “I²: Interactive Real-Time Visualization for Streaming Data” at the International Conference on Extending Database Technology (EDBT) together with his co-authors Nikolaas Steenbergen, Philipp Grulich, Tilmann Rabl and, Volker Markl. Due to its innovative solution on how to combine a data management system with an interactive user interface, the publication was presented with the Best Demonstration Award.
The publication describes a novel approach on how to visualize data streams. Because of the high data rates, it is oftentimes not possible to process all data points in a browser in real-time. As a result, a prior aggregation (data reduction) on a computer cluster is necessary. For this, I² combines the established open-source projects Apache Zeppelin (for visualization) and Apache Flink (for data stream processing) into an interactive development platform.
Next to the visualization of live data, the platform also allows for the interactive exploration of data streams. This ensures that possible visualization settings like filter criteria can be changed live. The current visualization settings are synchronized with cluster applications. The data preprocessing can therefore automatically adapt to changed settings with very low latency.
In their demonstration, the authors illustrated I² by the example of sensory data from a previously taped soccer game and demonstrated different interaction possibilities of the I² platform. The project is now available as open-source software on GitHub and DockerHub.
The committee honored the demonstration with the award because of its innovative approach of visualizing data streams as well as the interesting combination of data management systems with an attractive and interactive user interface.
