Stream Processing at Rovio: Where we are now / Where we would like to be at
This talk is about the journey from daily batch-based data processing towards streaming data at Rovio. We first do an overview of our analytics pipeline before going into more detail about the current applications of stream processing. The topics include real time monitoring, user anonymization, customization of user experience in games and attribution for user acquisition. There are also a number of areas where we should see benefits from using stream processing. We will briefly look at what needs to be solved for graph processing of identifiers, anomaly detection, machine learning and prediction (e.g, churning, lifetime value estimation and app install propensity) to get there.
Vesa Siivola is a Senior Data Engineer at Rovio Data Engineering Team in Espoo since 2016. He is both working with the operational side of the Rovio big data processing pipeline as well as developing solutions to derive insights from the data stream. He received his Ph.D. in 2007 from HUT (nowadays part of Aalto University) on language modeling for speech recognition.