Dark Data identification from event logs

Recently, process mining techniques were applied to discover the structure of big data pipelines from the event logs keeping track of their execution. To obtain further insights from the event log other than the traditional sequence flow based information and give value to the Dark Data involved in many pipeline executions, the objective of this thesis is to investigate some recent process mining solutions that exploit database theory to carry out analysis on the process behavior recorded over a relational DB. The student will build patterns of SQL queries to discover relations between the steps of a data pipeline that would be not made explicit by relying on traditional process discovery techniques.