Opportunities of Process Mining Research in
Software Engineering
Martin Macák
Faculty of Informatics, Masaryk University, Brno
About this presentation
∙ Aimed to provide a basic understanding of process mining.
∙ Overview of process mining applications in our lab.
Process mining
∙ Discipline proposed to give a better understanding of the processes by
extracting knowledge from event logs.
∙ Process: a series of related activities that are performed in a specific
sequence to achieve a particular goal within an organization.
∙ Event log: a structured collection of data that captures all the relevant
events that occur during the execution of a process.
Why is process mining important?
∙ We assume how the process is performed.
∙ However, how the real process looks like?
Process model
∙ A visual representation of a process.
∙ UML Activity diagram, BPMN, Petri net, . . .
Event log
∙ Case ID: a unique identifier
associated with a single
process instance.
∙ Activity: a specific step in
the process.
∙ Timestamp
Event log
id; activity; timestamp
23; order received; 2023-12-06 06:28:25
16; parcel picked up; 2023-12-06 07:12:38
23; depot transport; 2023-12-06 08:47:19
16; delivery started; 2023-12-06 08:50:06
41; order received; 2023-12-06 09:53:14
23; parcel prepared; 2023-12-06 10:02:11
Trace
∙ Trace is a representation containing only ordered activities of a given case.
∙ A collection of traces is called workflow log.
Event log
id; activity; timestamp
23; order received; 2023-12-06 06:28:25
16; parcel picked up; 2023-12-06 07:12:38
23; depot transport; 2023-12-06 08:47:19
16; delivery started; 2023-12-06 08:50:06
41; order received; 2023-12-06 09:53:14
23; parcel prepared; 2023-12-06 10:02:11
41; depot transport; 2023-12-06 10:07:42
41; parcel prepared; 2023-12-06 10:49:55
23;parcel picked up; 023-12-06 11:30:00
41;parcel picked up; 023-12-06 11:30:00
23; delivery started; 2023-12-06 12:10:16
<order received, depot transport,
parcel prepared, parcel picked up, delivery started>
<order received, depot transport,
parcel prepared, parcel picked up>
<parcel picked up, delivery started>
Workflow log
Process mining types
conformance
checking
process discovery
event log
event log + process model
process model
diagnostics
Process mining types
performance
mining
variants analysis
event log
event log event log
+
+
process model enhanced process model
diagnostics
Process discovery – Alpha algorithm
∙ Takes a workflow log.
∙ Outputs a Petri net.
∙ Petri net is a triplet N = (P, T, F) where:
∙ P is a finite set of places,
∙ T is a finite set of transitions such that 𝑃 ∩ 𝑇 = ∅,
∙ 𝐹 ⊆ (𝑃 × 𝑇) ∪ (𝑇 × 𝑃) is a set of directed arcs, called the flow relation.
ProcessM.NET
∙ For implementation of process mining algorithms, you can check our .NET
process mining library: https://github.com/lasaris/ProcessM.NET.
Declarative process mining
Declarative process mining
70
7
11
5
1
9
16
10
9
3
20
22
6
19
7
19
7
21
8
22
11
13
10
20
8
23
7
3
8
7
6
1
10
7
2
5
12
4
12
10
5
9
3
9
8
10
6
2
5
11
8
1
9
13
2
8
1
9
3
15
2
7
3
11
3
8
4
5
4
6
1
5
3
3
5
12
6
9
13
12
4
12
11
10
4
12
16
8
8
11
5
15
8
14
10
5
8
8
8
15
6
6
11
7
10
10
9
10
10
10
10
7
8
16
9
8
12
26
2
5
4
6
7
9
33
5
3
F
70
D
93
I
37
K
65
G
61
H
77
C
63
E
64
A
86
J
57
B
26
Declarative process mining
∙ If C occurs, then A occurs.
∙ Each time F occurs, then D occurs
immediately after F.
∙ B occurs only if preceded by C.
∙ Each time E occurs, then G occurs
afterward before E recurs.
∙ I and J never occur together.
♢(𝐶) → ♢(𝐴)
□(𝐹 → (𝐷))
(¬𝐵 ∪ 𝐶) ∨ □(¬𝐵)
□(𝐸 → (¬𝐸 ∪ 𝐺))
¬(♢(𝐼) ∧ ♢(𝐽))
Lasaris Process Mining Projects
∙ ERP processes
∙ Car sharing
∙ Software development
∙ Education
∙ Smartphone usage
ERP processes
∙ Cooperation with SAP Signavio
∙ process metrics
∙ process querying
∙ process atoms
Car sharing
∙ Cooperation with Autonapul and University of Tartu
Software Development
∙ Git logs
∙ ChatGPT utilization
Education
∙ Software development education
∙ Cybersecurity training sessions
Smartphone usage
∙ Cooperation with social scientists
Conclusion
∙ Process mining is fun!
∙ You can join the existing research or we can find a new area.