=Paper=
{{Paper
|id=Vol-2374/paper5
|storemode=property
|title=ProcessGold: Enterprise Process Mining
|pdfUrl=https://ceur-ws.org/Vol-2374/paper5.pdf
|volume=Vol-2374
|authors=Roeland Scheepens,Erik-Jan van der Linden
}}
==ProcessGold: Enterprise Process Mining==
https://ceur-ws.org/Vol-2374/paper5.pdf
ProcessGold: Enterprise Process Mining
1st Roeland Scheepens 2nd Erik-Jan van der Linden
Head of Research Non-Executive Director
ProcessGold ProcessGold
Eindhoven, The Netherlands Eindhoven, The Netherlands
roeland.scheepens@processgold.com erik-jan.vanderlinden@processgold.com
Abstract—The ProcessGold Enterprise Platform offers innovative or event sliders, apply different process mining algorithms,
features that are required for mature, scalable, and secure use filter, or select different case ID’s for multidimensional process
of Process Mining throughout large enterprises. Processes are mining, they are making different selections from the full event
displayed using process graphs, which require stable layouts to
ensure continued use by business users. The platform uses a novel data. Round trips to databases then hamper performance, and
graph layout algorithm, called Tracy, that ensures this stability hence in-memory architectures have become a standard feature
over user interaction. of enterprise process mining. Data selection, visualization, user
Relevance of enterprise features in general is evident from the navigation, as well as process mining algorithms all run in-
dispersion of process mining in enterprises, from use cases, and memory.
from broad adoption of the Platform.
Index Terms—Process mining, Enterprise process mining, sta- All features mentioned are part of the ProcessGold En-
ble graph layout terprise Platform. For a full overview of features see the
ProcessGold website [Pro] and sources below.
I. I NTRODUCTION
III. N OVEL FEATURE : T RACY PROCESS GRAPH LAYOUT
Since its inception around the turn of the century [AGL98],
the first decade of Process Mining [Aal16] saw the de- Besides these, by now common, features that meet require-
velopment from academic use and academic tools, such as ments for enterprise use, graph layout has become a topic
ProM [vDdMV∗ 05], to stand-alone tooling for use by individ- which requires attention to ensure process mining is contin-
ual analysts in businesses. The second decade, since around uously usable by business users. The ProcessGold Platform
2010, has been characterized by increasing use of and interest includes a novel algorithm for process graph layout, Tracy1 .
in process mining of non-technical business users in large A full technical paper about Tracy is available [MSW19].
enterprises.
A. Process layout
II. C OMMON FEATURES
Since its inception, Process Mining has mainly relied on exist-
The above development has led to a number of properties that ing algorithms for graph layout [GKN15]. However, these only
are by now common in Process Mining platforms. use the graph topology to compute a layout and do not make
Firstly, Process Mining tools have become part of the use of the particular properties of processes. For example,
system landscape of enterprises, which requires server- and most processes have some sort of a main path [RBRB06],
web-based architectures, and full compliance with standard [AEHK10] that is often the most frequent behavior. To fit
technical enterprise requirements such as data preparation on with intuitive semantics of process graphs, this path needs to
board, collaborative development, controlled deployment, and be centralized in the layout and depicted as a straight path
security. through the graph. While existing techniques fail to properly
Secondly, dispersion to non-technical business users re- do this [GKN15], the Tracy approach implements this.
quires easy point-and-click applications for everyday use, and
hence Process Mining has naturally integrated with Data Ana- B. Stable process graphs
lytics. This way, users have full access to required functionality Interaction with process graphs, required by business users,
from the field of Data Analytics. This includes any kind of by means of sliders, filtering, drill-down and alike leads to
business graphic besides process graphs, and full navigation changes in the layout of process graphs. Users build mental
options for selection, filtering, drill-down, etc. [Kei02]. maps of these graphs. When the graph changes, users need
Thirdly, for non-technical users, Process Mining is not their to fit their mental model with the altered graph. Conventional
primary task, and they require an optimal balance between algorithms [GKN15] lead to smaller or larger changes, and
time they spend working with the application on the one hand require users to at least put in significant cognitive effort
and business value on the other hand. In practice, users expect to adapt their mental map, but in the worst case even to
response times they experience in other web-based applica- rebuild their mental map altogether. Graph layout stability
tions. This performance requirement needs to be aligned with
the need for flexibility: when users apply, for example, edge 1 Patent pending.
A B
C D
Fig. 1. Graphs A and B show the same process as graphs C and D, respectively. The layouts in A and B are computed by the industry standard [GKN15],
where the actual process is poorly represented, while graphs C and D are computed by our novel graph layout algorithm where the process is easy to follow.
Graphs B and D are obtained after removing the edge highlighted in red from A and C. As we can see, B differs significantly from A; especially note how
nodes 1 and 2 swap vertically. Consequently, the mental map of the user is lost. On the other hand, C and D barely differ, preserving the mental map of the
user.
helps preserve the mental map of the user [PHG06], [ZKS11], Stability vs quality: Besides stability, we also require lay-
significantly reducing cognitive load. outs to be of high quality. Stability and quality are two
Therefore, we need to preserve the mental map when a conflicting requirements: graph layout stability helps preserve
new graph layout is displayed. The three models of Misue the mental map of the user, but also restricts the graph layout
[MELS95] to represent the mental map are our starting point algorithm in optimizing layout quality. A way of dealing with
to accomplish this. These models state that a layout adjustment this conflict is to allow somewhat larger changes to the layout
should preserve the direction of node n to node m for each pair and to make use of animation and transitioning as a secondary
of nodes n and m, that nodes that are close together should approach to mental map preservation. The combination of
remain close together, and that graphical objects in a region layout stability and transitioning provides the best approach
should stay in that region. to preserving the mental map.
In online dynamic graph drawing [BBDW17], graph layouts C. Graph layout: Tracy
need to be computed for a sequence of graphs without knowing
Tracy is a novel stable layout algorithm for process graphs
the full sequence from the beginning. In contrast, in offline
that computes layouts that intuitively represent the semantics
dynamic graph drawing [BBDW17], the whole sequence of
of the process. Our algorithm is based on the Sugiyama
graphs is known up front. Interactive Process Mining is offline
framework [STT81] but includes:
in the sense that we know which graphs we can potentially
encounter, since any filtered graph will be a subset of the data • A novel ranking algorithm;
found in the event log, but it is also online because we do not • A novel order constraint computation algorithm;
know the exact sequence of graphs beforehand. The Tracy- • A novel crossing minimization algorithm.
approach presents a novel way to take this into account, and In Figure 1, we present a comparison of Tracy and conven-
results in layouts that are stable for the user. tional layout.
Tracy is part of the production version of ProcessGold [AGL98] AGRAWAL R., G UNOPULOS D., L EYMANN F.: Mining pro-
starting from version 16, and will be publicly announced at cess models from workflow logs. In International Conference
on Extending Database Technology (1998), Springer, pp. 467–
ICPM 2019. 483.
[BBDW17] B ECK F., B URCH M., D IEHL S., W EISKOPF D.: A taxon-
IV. M ATURITY omy and survey of dynamic graph visualization. Computer
Graphics Forum 36, 1 (Jan. 2017), 133–159.
The ProcessGold Platform is a mature enterprise Process [GKN15] G ANSNER E. R., KOUTSOFIOS E., N ORTH S.: Drawing
Mining Platform. Customers of the platform include global graphs with dot. Tech. rep., Graphviz, Jan. 2015.
consulting firms, and global enterprises with over 1 billion [Kei02] K EIM D. A.: Information visualization and visual data
mining. IEEE transactions on Visualization and Computer
euro in annual turnover. Graphics 8, 1 (2002), 1–8.
A further indication of maturity is the business value that [MELS95] M ISUE K., E ADES P., L AI W., S UGIYAMA K.: Layout
customers get from implementing and using applications on adjustment and the mental map. Journal of Visual Languages
& Computing 6, 2 (June 1995), 183–210.
the platform. We present two use cases which illustrate this. [MSW19] M ENNENS R. J., S CHEEPENS R., W ESTENBERG M. A.: A
stable graph layout algorithm for processes. Eurographics
A. Use cases Conference on Visualization (EuroVis) (June 2019).
Telecom provider: A telecom provider is faced by increas- [PHG06] P URCHASE H. C., H OGGAN E., G ÖRG C.: How important
is the ”mental map”? – an empirical investigation of a
ing competition on price, requiring the need to reduce cost. dynamic graph layout algorithm. In Graph Drawing (Sept.
A process mining effort identifies and resolves a number 2006), Lecture Notes in Computer Science, Springer, Berlin,
of process issues, in particular related to manual exception Heidelberg, pp. 184–195.
[Pro] P ROCESS G OLD: ProcessGold - business intelligence and pro-
handling, communication with strategic suppliers. This leads cess mining platform. https://processgold.com/en/. Accessed:
to 20 percent personnel costs savings, and more broadly to 1+ 2019-23-02.
M cost savings in general, which equals 300 percent return on [RBRB06] R INDERLE S. B., B OBRIK R., R EICHERT M. U., BAUER T.:
Business process visualization – use cases, challenges, solu-
investment of the effort within 3 months. tions. In Proceedings of the Eighth International Conference
Insurance company: An insurance company is faced with on Enterprise Information Systems (ICEIS’06): Information
customers that are dissatisfied by untimely customer contacts. System Analysis and Specification (May 2006), INSTICC
PRESS.
Firstly, 71 percent of regular customer questions are not [STT81] S UGIYAMA K., TAGAWA S., T ODA M.: Methods for visual
responded to within the agreed service level, two days. A understanding of hierarchical system structures. IEEE Trans-
process mining effort identifies incorrect routing of questions, actions on Systems, Man, and Cybernetics 11, 2 (1981), 109–
125.
leading to delays. Once resolved, this dramatically improves [vDdMV∗ 05] VAN D ONGEN B. F., DE M EDEIROS A. K. A., V ERBEEK
the timeliness. Secondly, in case of disease of a person having H. M. W., W EIJTERS A. J. M. M., VAN DER A ALST W.
a pension insurance, the company witnesses highly negative M. P.: The prom framework: A new era in process mining
tool support. In Applications and Theory of Petri Nets 2005
comments from relatives on social media. These need to be (Berlin, Heidelberg, 2005), Ciardo G., Darondeau P., (Eds.),
prevented for their negative impact on the reputation of the Springer Berlin Heidelberg, pp. 444–454.
company. In the process mining effort, the 15 percent of cases [ZKS11] Z AMAN L., K ALRA A., S TUERZLINGER W.: The effect
of animation, dual view, difference layers, and relative re-
where communication is not timely handled are identified layout in hierarchical diagram differencing. In Proceedings
as well as the root causes, which are found in hand-over of Graphics Interface 2011 (GI’11) (2011), Canadian Human-
over cases and consulting secondary systems. By-catch is Computer Communications Society, pp. 183–190.
identification of 11 percent cost reduction.
B. Scale
Since 2016, customers and partners have acquired licenses for
use of the ProcessGold Platform for 10.000+ potential users.
ACKNOWLEDGMENT
Research and development presented in this demo paper was
carried out at ProcessGold. The authors acknowledge all
colleagues at ProcessGold that contributed to this work, as
well as academic partners at the Universities of Eindhoven (the
Netherlands) and RWTH Aachen (Germany) for productive
and long-lasting cooperation.
R EFERENCES
[Aal16] A ALST W. M. P. V. D .: Process Mining: Data Science in
Action. Springer, Apr. 2016.
[AEHK10] A LBRECHT B., E FFINGER P., H ELD M., K AUFMANN M.:
An automatic layout algorithm for BPEL processes. In
Proceedings of the 5th International Symposium on Software
Visualization (New York, NY, USA, 2010), SOFTVIS ’10,
ACM, pp. 173–182.