VisMillion: A novel interactive visualization technique for real-time big data

The rapid increase of connected devices causes more and more data to be generated and, in some cases, this data needs to be analyzed as it is received. As such, the challenge of presenting streaming data in such way that changes in the regular flow can be detected needs to be tackled, so that timely and informed decisions can be made. This requires users to be able to perceive the information being received in the moment in detail, while maintaining the context. In this paper, we propose VisMillion, a visualization technique for large amounts of streaming data, following the concept of graceful degradation. It is comprised of several different modules positioned side by side, corresponding to different contiguous time spans, from the last few seconds to a historical view of all data received in the stream so far. Data flows through each one from right to left and, the more recent the data, the more detailed it is presented. To this end, each module uses a different technique to aggregate and process information, with special care to ensure visual continuity between modules to facilitate the analysis. VisMillion was validated through a usability evaluation with 21 participants, as well as performance tests. Results show that it fulfills its objective, successfully aiding users to detect changes, patterns and anomalies in the information being received.