Multi-Dimensional Process Mining using Neo4j and Cypher

Process mining, so far, has required sophisticated, special-purpose software to handle, filter, analyze event logs, discover models, and analyze deviations. This sequence of tutorials shows how to use a general purpose graph database system (Neo4j) and graph query language Cypher for process mining. The tutorials shows how to use Neo4j and Cypher to

  1. load event data into Neo4j, specifically event data over multiple case/entity identifiers
  2. how to construct and extend event knowledge graphs from event data over multiple case/entity identifiers with simple Cypher queries
  3. how to analyze event knowledge graphs through querying and filtering
  4. how to discover basic process models through aggregation directly in event knowledge graphs
  5. how to detect task execution patterns and construct a multi-layered event knowledge graph for advanced analyses over multiple levels of abstraction

The ready-made query templates enable users to begin learning process mining over multiple entities and in multiple behavioral dimensions and to design their own analysis stacks with very low effort: by simply using off-the-shelf graph database systems and standard query languages.

How to install:

  1. Download and install Neo4j desktop
  2. Download the latest release (.zip) of the tutorial, including query templates Python scripts, and example data from
  3. Unzip into a local directory (on a path without dash character ‘-‘)
  4. Open ./order_process/ (or .pdf)

Example data sets:

Contact person providing support during the summer school: Dirk Fahland (

Providing support on the following days: Monday evening (4th July 2022) through Friday noon (8th July 2022)