Process Mining: closing the gap between Data Science and BPM

Process Mining
Closing the gap between Data Science and BPM
Conferencia de Posgrado
Kybele
Research
Group
Juan Manuel Vara [Juancho]
juanmanuel.vara@urjc.es
@KybeleResearch | @jmvara
Facultad
de
Informática

Intro
This aims to be a very applied talk
• The focus is not (only) to introduce the scientific basis of Process Mining
• But to show the potential of the topic
• Foster interest
• To be aware that practical application is not only feasible but almost direct
• To go deep into its application to your field of study
• To have a first glance at the theory behind
Process Mining | @jmvara 2

The speaker
The beginning
• Ingeniería Aeronáutica - UCM
• IngenieríaTécnica enTelecomunicaciones - UPM
• Ciencias Ambientales - URJC
• Ingeniería Informática – URJC
• IBERIA –VivaTours

The speaker
Life at @URJC
• Head of the MsC in Information Systems Engineering
• Head of the SE, Informations Systems and Service Engineering at the PhD Program onTIC
• Head of the BsC in Services Engineering

The speaker
My life as a researcher
• Pre-doc at U. Nantes (Jean Bézivin)
• Post-doc atTilburg University (Mike Papazoglou)
• 2 Sexenios / 1Transferencia
• 4 Docentia
• 7Tesis Doctorales en los últimos 10 años
• 18 artículos JCR – 16 congresos CORE
• IP proyectos MINECO
• Investigador H2020
• …
Model
Driven
Engineering

Agenda
Motivation
What is in Process Mining?
Quick glance at Data Handling
A glance at Process Discovery
Quick tour with Disco
• Process Mining tools overview
Challenges & Open Issues

Motivation
7
Process Mining | @jmvara

Business Process Management
One straightforward way of looking at BPM is to think of 10W’s associated with a business
Business Process Management (BPM) is the art
and science of overseeing how work is performed
in an organization to ensure consistent outcomes
and to take advantage of improvement
opportunities.
By understanding WHAT is done, WHY, WHERE and WHEN it is done, WHO
does it, the WAY it is done, and WITH what resources, all while keeping a
WATCH on performance, identifying what to WORRY about and how to mitigate
those worries, a range of methods and tools can be applied to achieve the best
results with the least cost and achieving the most WEALTH.
https://www.igrafx.com/articles/why-business-process-management/
A collection of discrete
activities or events we
perform in order to achieve a
certain goal

BPM Issues
Creation of “current state” processes
• Business Process reengineering focus on an improved to-be process
• Little interest in exploring the as-is one, which is the key to:
• Know whether it’s worh to invest in improvements
• Detect performance problems
• How much variation there is in the process across the organization
• Skip or subcontract process analysis
• Spend too much time in interviews and sticky notes based process analysis
Lack of connections between BPs and ISs
• Gathering information about how your process is performing entailed manual and complex steps to collect and
synthetize data
• E.g.: Lean & Six Sigma have not emphasized technologies as enables of Process Management
(Davenport & Spanyi, 2019)
80/20
• 80% of the process instances are
described by 20% of the variants
• The remaing 20% of process
instances make up for 80% of the
problems (deviations, delays,
complaints, etc.)
• Data-based as-is model
• BPM initatives connected
to ISs

The promise of RPA
Robotic Process Automation
• APIfication
• Incorporación de ML: cognitive RPA
• Low-code tools
Companies will flesh out their stacks by combining
RPA, traditional BPM, workflow, and cloud
services, layering in process documentation,
process discovery and analytics, and looking to
enable a ‘virtuous cycle’ of process improvement.
Ryan Duguid, Nentix. From: RPA (Robotic Process Automation):
What’s In Store For 2020?. Forbes, Dec.2019
• Identify where to implement bots
• Calculate the impact of the
implementation

The Hype of low-code
The rise of low-code & no-code tools
• Comunity or Citizen Developers
• Low-code for pro-developers
Sortable list of low-code platforms
https://www.managedfunctions.com/pirate
(Bridgewater, 2021)

Our interest on Service Design
TO-BE AS-IS
How do we know
what was really
going on there?
How do we
know what is
really going on
after our
intervention?
INNoVaServ: model-
based tool support for
Service Design
https://modeling-languages.com/model-based-tool-for-business-service-design/
Pérez-Blanco, F. J., Vara, J. M., Gómez, C., De Castro, V., & Marcos, E. (2020,
April). Model-based tool support for Service Design.
In International Conference on Fundamental Approaches to
Software Engineering (pp. 266-272). Springer, Cham.
Gómez, C., Pérez Blanco, F. J., Vara, J. M., De Castro, V., & Marcos, E. (2021,
January). Design and development of Smart Contracts for E-
government through Value and Business Process Modeling.
In Proceedings of the 54th Hawaii International Conference on
System Sciences (p. 2069).

No one wants to
ends like this
Process Mining
Data-driven approach to
discovering gaps in your
business process

Process Mining Overview
Adapted
from
@mathias_weske:
https://twitter.com/mathias_weske/status/1388036059871784962?s=20

A more pragmatical contribution
I’m personally not sure about the enhacement issue
More prone to think about
Throughput analysis/bottleneck
detection
Accounting for the intensity of events’ execution (measured by time
spent to complete a particular event) in order to determine potential
bottlenecks.
This kind of analysis can be used to improve time-related KPIs by
minimizing throughput/overhead time

What is in Process
Mining?
22

A bit of history
One man’s Dream
• Wil van der Aalst dissapointment on the gap
between modelled processes and reality
• Late 90’s sabatical at U. of Colorado in Boulder
• Early 2000s the 𝝰-algorithm
• 2004: first ProM release
• 2007: foundation of Futura PI (van de Brand)
• Fluxicon, Celonis, ProcessGold
• 2014: Coursera Process Mining course (120k participants)
• 2015: practical adoption of Process Mining in the industry
(Van der Aalst, 2020)

What is Process Mining?
Process Mining
• Lies at the intersection of process-science and data-science, bridging the gap between model-based process analytics
and data-centered analytics.
• It allow users to reconstruct, analyze and improve business processes based on event logs from transactional IT
systems like SAP, Oracle, Salesforce, etc.
• Undesired process patterns, bottlenecks and compliance issues can be detected and reduced in that way. This allows
us to understand processes as they really happened based on their digital footprint.
• This reflects a major advantage to previous methods of process mining and
modeling as findings are no longer based o assumptions but on
data-based evidence reflecting real-world events.
Adapted from (van Der Aalst, 2016)
Process Science
• Process Management
• Automation
• Process Control
• Process Improvement
• Operations
Management
Data Science
• Data Mining
• Satistics
• Machine Learning
• Databases
• Predictive Analytics
Process
Mining
The end goal of process mining is
to discover, model, monitor, and
optimize the underlying processes
by analyzing the event data,
generated during the execution of
those processes.

What is in Process Mining
It’s basically about analysing digital
footprints, don’t matter what are they
IT-based work
Every workflow is supported
by IT systems, like SAP,
Oracle or Salesforce
Digital Footprints
Process Minig algorithms detect &
reconstruct workflow traces
Full Transparency
Actual process executions are
visualized in the tool

What do we need for Process Mining
Data generation
• Business process data is stored in a variety of IT systems, suh as
workflow management systems, ERP systems, CRM systems,
supply chain management systems, etc.
Celonis (2020)
Event logs
• Store the data that is required for Process Mining.
• At the minimum, the event log has to include three
different fields:
• Case ID
• Activity Name
• Timestamp
• There may be other optional fields
Adapted from: van Der Aalst (2016)
Digital Footprint

Digital Footprints
• Digital Footprint (key parts of an event)
• Case ID: indicates which process instance the event belongs to.
• A unique identifier of an entity going through the process
• Activity / Event: describes the action that is captured by the event
• A step of the process, any activity that is a part of the process
we are analyzing
• Timestamp: indicates the time when the event took place
• used for performance evaluation and determining the order of
events, can be the time when the user entered/exited the given
event (or both actually)
By including additional
information (resources, country,
user segment, etc) we are able to
carry out a much more detailed
analysis.

Digital Footprints
#123
12-02-21
Receive
Order
#123
18-02-21
Create
Invoice
#123
16-02-21
Ship
Goods #123
23-02-21
Receive
Payment
#ID
Time
Activity

Cases &Variants
• Case / Trace: a particular sequence of activities
• A case usually consists of multiple events
• Variant: the set of cases/traces that perform the very same
sequence of activities ≅ Process Pattern

Quick glance
(Data import)
32

How does it work?
Case Activity
1 A
2 A
1 B
1 C
3 A
2 C
3 B
2 B
1 D
2 D
2 E
3 C
3 D
1 E
3 D
3 E
A
B
C
D
E
Case 1 Case Activity
1 A
2 A
1 B
1 C
3 A
2 C
3 B
2 B
1 D
2 D
2 E
3 C
3 D
1 E
3 D
3 E
A
B
C
D
E
Case 1
A
C
B
D
E
Case 2 Case Activity
1 A
2 A
1 B
1 C
3 A
2 C
3 B
2 B
1 D
2 D
2 E
3 C
3 D
1 E
3 D
3 E
A
B
C
D
E
Case 1
A
C
B
D
E
Case 2
A
B
C
D
E
Case 3
D
https://fluxicon.com/book/read/intro/#how-does-it-work

How does it work?
Case Activity
1 A
2 A
1 B
1 C
3 A
2 C
3 B
2 B
1 D
2 D
2 E
3 C
3 D
1 E
3 D
3 E
A
B
C
D
E
Case 1
A
C
B
D
E
Case 2
A
B
C
D
E
Case 3
D
A
B
C
D
E

How does it work?
Case Activity
1 A
2 A
1 B
1 C
3 A
2 C
3 B
2 B
1 D
2 D
2 E
3 C
3 D
1 E
3 D
3 E
A
B
C
D
E
Case 1
A
C
B
D
E
Case 2
A
B
C
D
E
Case 3
D
A
B
C
D
E
Case Activity
1 A
2 A
1 B
1 C
3 A
2 C
3 B
2 B
1 D
2 D
2 E
3 C
3 D
1 E
3 D
3 E
A
B
C
D
E
Case 1
A
C
B
D
E
Case 2
A
B
C
D
E
Case 3
D
A
B C
D
E

How does it work?
Case Activity
1 A
2 A
1 B
1 C
3 A
2 C
3 B
2 B
1 D
2 D
2 E
3 C
3 D
1 E
3 D
3 E
A
B
C
D
E
Case 1
A
C
B
D
E
Case 2
A
B
C
D
E
Case 3
D
A
B
C
D
E
Case Activity
1 A
2 A
1 B
1 C
3 A
2 C
3 B
2 B
1 D
2 D
2 E
3 C
3 D
1 E
3 D
3 E
A
B
C
D
E
Case 1
A
C
B
D
E
Case 2
A
B
C
D
E
Case 3
D
A
B C
D
E
Case Activity
1 A
2 A
1 B
1 C
3 A
2 C
3 B
2 B
1 D
2 D
2 E
3 C
3 D
1 E
3 D
3 E
A
B
C
D
E
Case 1
A
C
B
D
E
Case 2
A
B
C
D
E
Case 3
D
A
B C
D
E

A glance at
process discovery
40

The Alpha Miner
Algorithm
• Scan traces for ordering relations
• Build footprint matrix
• Convert footprint matrix to Petri Net

The Alpha Miner
Build footprint matrix
• Four ordering relations
• > Directly follows a > b a is directly followed by b
• → Sequence a → b if a > b and not b > a
• || Parallel a || b if both a > b and b > a
• # Choice a # b if neither a > b and b > a
L1 =[⟨a, b, c, d⟩3,⟨a, c, b, d⟩2,⟨a, e, d⟩]
⟨a, b, c, d⟩
⟨a, b, c, d⟩
⟨a, b, c, d⟩
⟨a, c, b, d⟩
⟨a, c, b, d⟩
⟨a, e, d⟩

The Alpha Miner
Convert footprint matrix to Petri Net
L1 =[⟨a, b, c, d⟩3,⟨a, c, b, d⟩2,⟨a, e, d⟩]
⟨a, b, c, d⟩
⟨a, b, c, d⟩
⟨a, b, c, d⟩
⟨a, c, b, d⟩
⟨a, c, b, d⟩
⟨a, e, d⟩

The Alpha Miner
L1 =[⟨a, b, c, d⟩3,⟨a, c, b, d⟩2,⟨a, e, d⟩]
⟨a, b, c, d⟩
⟨a, b, c, d⟩
⟨a, b, c, d⟩
⟨a, c, b, d⟩
⟨a, c, b, d⟩
⟨a, e, d⟩

The Alpha Miner
Limitations
• Implicit places: places that can be removed without changing the behavior
• Harmless, but the model becomes unnecessarily complicated.
• Representational bias
• Mixture of choice and synchronization
• Does not deal well with loops of length 1 and 2
L =[⟨a, c⟩2, ⟨a, b, c⟩3, ⟨a, b, b, c⟩2, ⟨a, b, b, b, b, c⟩1] L =[⟨a, b, d⟩3, ⟨a, b, c, b, d⟩2, ⟨a, b, c, b, c, b, d⟩]

The Alpha Miner
Limitations
• Non-local dependencies
• Dependency between two activities that never follow
one another directly
• Frequencies are not taken into account.
• Very sensitive to noise and incompleteness
• Foundational problems for process mining
• Don’t have negative examples
• Soundness
• Safeness: places cannot hold multiple tokens at the same time
• Proper completion: if the sink place is marked, all other places are empty
• Option to complete: it is always posible to reach the final marking
• Absence of dead parts: for any transitions, there is firing sequnce enabling it
L =[⟨a,c,d⟩45,⟨b,c,e⟩42]
L‘= [⟨a, c, d⟩45, ⟨b, c, d⟩42, ⟨a, c, e⟩38, ⟨b, c, e⟩22]
Underfitting
Desired Model

Other discovery approaches
Heuristic mining
• Similar representation to Causal Nets
• Focus on dealing with noise & incompleteness
• Take frequencies and sequences into account
• Avoid incorporating infrequent paths in the model
• Non-fitting model
Fuzzy mining
Genetic process mining
Region-based mining
• Able to express more complex control-flow structures
without underfitting
Inductive mining
• Sound models & highly scalable
(van der Aalst, 2019)

Quick tour with Disco
a simple yet powerful tool for Process Mining
53
Adapted from (Fluxicon, 2020)

Example Scenario
Purchasing process
Requester Purchasing
Agent
Requester
Manager
Supplier Financial
Manager

Objectives
1. Inefficient operations
2. Need to demonstrate
compliance
3. Complaints about process
duration
1. Understand the process
in detail
2. Control performance
targets (21 days)
3. Check whether there are
deviations from the
payment guidelines
Problems Goals

Process Mining Projects Lifecycle - Roadmap
Questions
• Determine questions
• Process scope
• Which IT systems
Data
Extraction
• Via IT staff
• CSV file or database
extract
Data Analysis
• Discover ‘As-is’
process
• Answer questions
Presentation
• Present results
(workshop, report,
etc.)

Roadmap - Questions
1. Understand the process in detail
2. Control performance targets (21 days)
3. Check whether there are deviations from
the payment guidelines
1. How does the process actually look like?
2. Do we meet the performance targets?
3. Are there deviations from the prescribed
process?
Questions
• Process scope
Data
Extraction
• Via IT staff
extract
Data Analysis
process
Presentation
• Present results
(workshop, report,
etc.)

Roadmap – Data Extraction
Questions
• Process scope
Data
Extraction
• Via IT staff
extract
Data Analysis
process
Presentation
• Present results
(workshop, report,
etc.)
IT staff extracts history logs from the
ERP system

Things to consider for your data extraction
Which process
• Availability of data
• Champion support
• A sponsor who wants a “surprise me” analysis is a red flag
• Potential for Improvement
RQ
• Focus your analysis
• Pose aditional requirements for data extraction
IT Systems
• CRM, ERP, SAP
• Workflow, ticketing systems, ad-hoc systems
Case ID
• Combination from multiple fields
• Different Case IDs in different parts of the process
• N:M Relationships
Different Case IDs in different parts of
the process
1. Sales order: traced by Sales order ID
2. Delivery: traced by Delivery ID
3. Invoicing: traced by Invoicing ID
N:M Relationships
“flattening reality” (like
putting a 3D-world in a 2D-
picture).
• You need to choose
which perspective you
want to take on your
process.

XES (eXtensible Event Stream)
Official IEEE Standard
• Driven by IEEETask Force on Process Mining
• XML syntax and OpenXES library
• Supported by most of the existing tools
• Conversion from other formats (CSV) is easy
www.xes-standard.org
(Weske,
2021)

Questions
• Process scope
Data
Extraction
• Via IT staff
extract
Data Analysis
process
Presentation
• Present results
(workshop, report,
etc.)
Roadmap – Data Analysis
1. Inspect Data
2. Import Data
3. Inspect Process
4. Inspect Statistics
5. Inspects Cases
6. Filter on performance
7. Visualize Bottlenecks
8. Animate Process
9. Compliance Check
10. Organizational View

1. Inspect Data
Case 1
Event

2. Import Data
Check the
guess
made by
the tool

3. Inspect Process
• Numbers in rectangles
are activity
frequencies
• All 608 cases start
with activity ‘Create
Purchase
Requisition’
• Number at arcs is
frequency of
connection
Numbers, thickess and
coloring indicates frequency

3. Inspect Process
• Lots of changes
were made
(amendments)!

3. Inspect Process
• At the lowest point only the
activities of the most
frequent variant (Happy
Path) are shown
Happy
Path

3. Inspect Process
¿?
100%

3. Inspect Process
+
100%

3. Inspect Process
+
100%
Why seeing less tan 100% of
your process?
‘Cause otherwise your process
map might be too complicated
to derive any useful insights or
actionable information from it.

Timeframe
Data Set

Timeframe
Data Set
Where in the process are
are spending so much time?

5. Inspect Cases
Variants ordered by
frequency: 98 variants,
the most frequent
comprising 88 cases
that make up for 15% of
the data set
10% of all
cases follows
pattern 3

5. Inspect Cases
History of each
particular variant
and/or case:
enables root-cause
analysis
Why are so many
requests stopped?
Do people not know
what they can buy?
Early
end-point

Results so far
1.How does the process actually look like?
• Objective process map discovered
• Lots of amendments and stopped requests:
Update of purchasing guidelines needed
2.Do we meet the performance targets?
• Not by all (some take longer than 21 days):
Where in the process do we lose the time? à Next
3.Are there deviations from the prescribed process?

Mind your steps:
we are not
looking at the full
data set now

On average 3
amendments per
case If you want to see how
many cases go through that loop,
then you can switch from
Absolute frequency to Case
frequency in the drop-down list

7.Visualize bottlenecks
Total durations
(the cumulative
delays over all
cases) shows the
high-impact areas
Mean durations
let us inspect
average delays
On average it
takes more
than 14 days
to return from
the rework
loop to the
normal process

7.Visualize bottlenecks
Process mining show us where we have problems
• Data-based analysis allow us to objectively know
where we should focus
• Waiting times are often magnitudes higher than
execution times
• Focus is not on making people work faster but
to organize the process in smarter way
Process mining helps to communicate
our findings
• Change initiatives are hard
• Processes are complex to understand
• Charts and statistics are ofetn too abstract
• Visual representations might help to engage
stakeholders
On average it
takes more
than 14 days
to return from
the rework
loop to the
normal process
Interactive analysis workshops to
take advantage from domain
experts

8.Animate Process
Animation can be extremely helpful
for communication
• Can make the discovered bottlenecks
really tangible for people
and “bring them to life”
Dynamic
replay of the
process over
the time frame
of data that we
have
Cases moving as its
actual, relative speed
Drag needle to end
of time line and
observe how the
most used paths get
thicker and thicker

Results so far
• Not by all (some take longer than 21 days)
• The ‘Analyze Request for Quotation’ activity is a huge bottleneck:
Process change is needed
3. Are there deviations from the prescribed process? à Next

9. Compliance Check
10 (out of 608) cases
skip the mandatory
‘Release Supplier’s
Invoice’ activity!

9. Compliance Check
We can filter
the data set to
focus on the
cases that go
through a
particular path
Actionable result:
• Change the
operational system
to prevent the
violation
• Provide targeted
training

Results so far
• Not by all (some take longer than 21 days)
• The ‘Analyze Request for Quotation’ activity is a huge bottleneck:
Process change is needed
3. Are there deviations from the prescribed process?
• Yes, training or system change needed

10. OrganizationalView
Alternative views on the data
• Re-configure the data import

10. OrganizationalView
Different perspective
• We can take many different perspectives, even
based on the same data set
We now don’t see the activity sequences displayed
but rather the hand-over of work between
different functions, or departments
• Inefficiencies can often be found at the borders of
organizational units
Clearly, the
Purchasing
agents are
causing the
biggest delays
in the process!

Take-away
A few things you can learn
• Real processes are often more complex than you might expect
• Manual process analysis get the Sunny day scenario
• Process Mining get a hold of all the exceptions
• There is not one right model
• Process mining is not about mining a data set to create one process model
• Explore the process by translating questions into filters
• Interactive activity needing from domain knowledge to interpret findings

The hype

Tooling landscape
ProM
• Open-source, extensible framework
• More than 1500 plugins (#29 in 2004!)
Over 35 process
mining vendors by
the end of 2020
• Lower the threshold for process mining.
• Inability to discover concurrency well.
• Focus on performance analysis rather than
conformance checking and precise models.
CELONIS
• Cloud-based + BI capabilities
• Strong connection with SAP
Disco
• Low threshold
• Easy to use yet scalable

Event
Collection
Data Pool
Data
Connection
Data Job
Extractions
Extraction Task
---
Transformations
Transformation
Task
---
Data Job
Data Job
Data
Connection
Data
Connection
Data Pool
---
CELONIS
Much more powerful and rich environment
• Combined BI and Process Mining approach
• Wide collection of data engneering options
• Off-the-shelf components (e.g. SAP)
Additional Complexity
Separate
Database
Database
Schema

Data Engineer Tasks
Create new Data Pool in Event Collection
Create Data Connections to your source systems
Create a Data Job and set up an Extraction Task to extract all relevant tables
Set up the Transformation Tasks
• Create ActivityTable
• Insert activities into the empty ActivityTable
• Specify object
• Select Informacion
• Insert Activity
Create a data model and extend it by adding master and data tables
Apply name mapping
Connect
Extract
Transform
Set up
data model

Challenges &
Open Issues
92

Directly-Followed Graphs Issues
Existing tools start from DFGs for Discovery
• DFGs cannot express choice & concurrency
• This may lead to Spaghetti-like DFGs with loops
• Filtering DFGs using frequency-based thresholds
may provide misleading results
• Performance information mapped onto DFGs can be misleading
• The average time reported between two activities is conditional
• Limited support for conformance checking
• Rule-based approach being the least bad solution
Simplicity & Performance
(over accuracy)

Hurdles for widespread adoption
Data quality
• 80% of the efforts and time are spent on locating, selecting, extracting and transforming the data
• Time needed to apply process mining is short (say 20%) once the data is available in the right format
People
• Stakeholders are unaware of process mining capabilities
• Hard to distinguish from ML and AI
• May reveal mismanagement and complicance problems
(Gartner, 2020)

Trends (mainly Academic)
DigitalTwins of an Organization (DTO)
Process Discovery as the starting point
• Uptake in conformance checking & performance analysis techniques
• Data Mining & ML to find root causes
“A dynamic software model of any organization, that relies on
operational and/or other data to understand how an
organization operationalizes its business model, connects
with its current state, responds to changes, deploys resources
and delivers expected customer value.”
(Kerremans, 2018)
None of the Use Cases compiled in
(Reinkenmeyer, 2020) dealing with
conformance checking

Trends (mainly Academic)
Forward looking
• Backward-looking can be used to fundamentally improve processes
• Provides little support for the day-to-day management of process
• Need to be able to analyze cases that are still running
Call to action
• Process Mining to identify candidates for RPA and monitor SW robots
• RPM to translate Process Mining diagnostics into management actions
(van der Aalst, 2020)

The Challenge
The need for Business Process Hygiene
From silos to corporate process mining
• The need for a digital workforce
• Mindset change
• “do I have to use data” à “can I afford not to use available data”
We are probably paving the way
• Spreadsheets for process analysts
• 35+ Process Mining tools
• More than 135K following the
Process Mining MOOC
(van der Aalst, 2019; Reinkenmeyer, 2020b)
There exists so much
information that people are
looking for new ways to
leverage their data

Personal reflection
Academia
Basic Research
Algorithms Discovery
Conformance Checking
Predictive process monitoring
Object-centric Process Mining
Data Engineering Automate ETL
Deal with uncertain and continuous
event data
Industry
Applied Research
New application fields University Processes Portfolio
Service Design
Blockchain
New applications in explored fields Gather new insights
Ad-hoc methods & techniques

References
• Bridgewater,A. (2021).A RisingTide, HighTimes For Low-Code. Forbes, Jan 27, 2021.
• Celonis.(2020). Process Mining for Experts Certificate.
• Davenport,T. H., & Spanyi,A. (2019).What process mining is, and why companies should do it. Harvard Business
Review, 97(2).
• Fluxicon. (2020). Hands-on Tutorial. https://fluxicon.com/book/read/tutorial/
• Gartner. (2020). Interview with Prof.dr.ir.Wil van der Aaalst. In 2020 Gartner Market Guide for Process Mining.
Research Note G00733123
• Kerremans, M. (2018). Market Guide forTechnologies Supporting a DTO. Gartner Inc.
• Jans, M. (2017). From relational database to valuable event logs for process mining purposes: a procedure.Technical report,
Hasselt University.
• Reinkemeyer, L. (2020). Process Mining in Action. Cham: Springer International Publishing.
• Reinkemeyer, L. (2020b). BusinessView:Towards a Digital Enabled Organization. In Process Mining in Action (pp. 197-
206). Springer, Cham.
• Rozinat,A., Günther, C. (2018). Disco, Process Mining for Professionals. : https://fluxicon.com/

References
• van der Aalst,W. (2016). Process Mining: Data Science in Action. Springer, Berlin, Heidleberg.
• van der Aalst,W. (2018). (MOOC) Process Mining: Data science in Action. Coursera:
https://www.coursera.org/learn/process-mining
• van der Aalst,W. (2019). Business Process Hygiene: No Business Case Needed. https://bit.ly/3fg2KpR
• van der Aalst,W. (2020).Academic view: Development of the process mining discipline. In Process Mining in Action (pp.
181-196). Springer, Cham.
• Weske, M. (2021). Process Mining:An openHPI course. Practical Aspects Simulating Event Logs.

Process Mining: closing the gap between Data Science and BPM

More Related Content

Similar to Process Mining: closing the gap between Data Science and BPM (20)

More from Facultad de Informática UCM (20)

Recently uploaded (20)

Process Mining: closing the gap between Data Science and BPM