[gota]

Process Mining [1]. When I programmed a rather complex logistics simulator at work I told my coworkers 'whoever comes up with a way of instantiating a simulator from data will be praised forever'. Turned out process discovery is a thing (well, one of _the_ things in PM). And there's so much cool stuff to do and being done. I'm now on the last stretch of my doctorate researching the mining of typical plans in non-competitive environments.

[1] https://en.wikipedia.org/wiki/Process_mining

[janemanos]

Have you ever had a look into analyzing processes with a graph database like e.g. ArangoDB. Wonder if that would make sense for your needs. You can traverse along the processes, find patterns or use distributed graph processing with Pregel analyze from different angels. edit:typo

[whazor]

It is not about the big scale of processes that make process mining interesting. But also the tooling that comes with the field, look for example at the tool Disco: https://www.youtube.com/watch?v=pmXZQhFSv10

It provides automatic visualisation of graphs, analysing of bottlenecks, and lots of analytics. While you only need system logs linked to an id.

[gota]

I haven't, no, but I know there are initiatives in Process Mining that closely relate to knowledge graphs, etc. and wouldn't be surprised if there are groups working on that.

I'm particularly working with the mining of plans (as in Automated Planning) in declarative process models. If I have a chance I'll look into it, thanks for the heads up