Process mining in
IS MU
Jakub Hanko & Daniela Krúželová
File categories
• Scripts (3891)
• Modules (722)
Hypotheses to
be verified
• Summer time vs. rest of the year
• .pm (modules) vs. .pl or .js (scripts)
Our typical
activities in Git
• Big change (>200)
• Medium change
• Small change (<10)
• Rename
• Revert
• Quick fix
What were our
steps?
• Discovering typical activities in our git
repository
• Writing perl script to generate data into
csv
• PM discovery by Disco
Our discoveries
• Summer time:
o less activity in modules
o average number of committed changes is
higher in modules (new agendas)
o more in Disco
Our discoveries
• Scripts vs modules:
oIn both cases small and medium commits
were predominant
oRatio of small to medium commits in
modules is almost 2 : 3 but in scripts it is
2 : 5
oQuick fix and reverts usually don’t
happen after a big commit
oBig commits were more often in modules
oLess renames in modules
Our conclusions