Filter Large-Scale Engine Data Using Apache Spark

This paper introduces a minimum viable software product to filter large datasets of engine data recorded during laboratory experiments of combustion engines. The aim is to support analysts in the identification and analysis of specific physical phenomenon within hours of recorded engine experimental data. Specifically, the tool has been designed considering the use case of identifying Low Speed Pre-Ignition events. This work describes the tool's graphical user interface and its scalable architecture based on mainstream web and big-data technologies as well as the practical application to pre-ignition events identification. The paper provides details on the architecture's performance, providing evidence of its scalability by increasing the number of available computing workers.