Most business processes are supported nowadays by information systems, that record event data about the executions of the processes. Process mining plays an important role in linking the BPM field with data science, helping to identify the bottlenecks and the unwanted behavior, and to adopt strategies to improve the process, measuring the eventual benefit. While many algorithmic techniques have been developed for discovery, conformance checking and other process mining techniques, extracting

more »

... a from today's information systems requires the specification of a complex query that extracts the required information and groups the events in cases. The research project described in this paper proposes a novel framework to support process mining analysis, that uses the advances in graph algorithms and in-memory data processing in order to reduce the costs of extraction and transformation of the event data contained in the information systems. At the end of the project, a set of pre-processing, discovery and conformance checking techniques, that do not require the specification of a case notion, will be made available in different environments, technologies and languages, e.g., PM4Py, Spark, Neo4J, Celonis. In comparison to related work, this research aims to obtain a complete and scalable framework that supports process mining from the Extraction, Transformation and Load phase (from relational and non-relational databases) to the effective analysis/usage of the data, and to get a class of process models fully capturing the lifecycle and the interactions between different classes. Since the framework is aimed at real-life, complex, information systems, a goal of the project is to attain significantly better scalability than existing approaches.