=Paper=
{{Paper
|id=Vol-2839/paper6
|storemode=property
|title=A Dashboard-based Approach for Monitoring Object-Aware Processes
|pdfUrl=https://ceur-ws.org/Vol-2839/paper6.pdf
|volume=Vol-2839
|authors=Marius Breitmayer,Lisa Arnold,Manfred Reichert
|dblpUrl=https://dblp.org/rec/conf/zeus/HeinzeAS21a
}}
==A Dashboard-based Approach for Monitoring Object-Aware Processes==
https://ceur-ws.org/Vol-2839/paper6.pdf
A Dashboard-based Approach for Monitoring
Object-Aware Processes
Marius Breitmayer, Lisa Arnold and Manfred Reichert
Institute of Databases and Information Systems, Ulm University, Germany
{marius.breitmayer,lisa.arnold,manfred.reichert}@uni-ulm.de
Abstract. Data (e.g., event logs) gathered during the execution of busi-
ness processes enable valuable insights into actual process performance.
To leverage this knowledge, these data should be analyzed and inter-
preted in the context of the respective processes. Corresponding analy-
ses allow for a comprehensive process monitoring as well as the discovery
of weaknesses and potential process improvements. This also applies to
object-aware processes, where data drives process execution and, thus,
is treated as first-class citizen. This paper introduces a dashboard with
advanced monitoring functions for object-aware processes.
Keywords: object-aware processes, process monitoring, dashboard
1 Introduction
Process monitoring leverages the data created during the execution of business
processes in order to gain insights into actual process performance and to ensure
process conformance with regulations, policies, and business rules [12]. This in-
formation, in turn, can then be used to monitor, analyze and improve processes
[8]. Despite existing approaches for monitoring activity-centric processes (e.g.,
[5]), adequate monitoring support for data-centric paradigms to business process
management (BPM) [16], e.g., artifact-centric processes [7], object-aware pro-
cesses [10], and case handling [2], is still lacking. As opposed to activity-centric
processes, in object-aware process management, large process structures involv-
ing multiple concurrently executed and interacting object lifecycle processes,
emerge during runtime [15,4]. The monitoring of such dynamically evolving pro-
cess structures constitutes a particular challenge not sufficiently addressed by
contemporary approaches [14]. As data is treated as first-class citizen in object-
aware process management, however, the monitoring of data-centric processes
yields great potential for more advanced analyses and better comprehensibility
of processes in general due to the tight integration of process and data.
To set a background, a short introduction of object-aware process manage-
ment is provided. PHILharmonicFlows, our approach to data-centric BPM [10],
introduces the concepts of objects, object behavior, and object interaction. Each
business object of a real-world business process is represented as one such object.
The latter comprises data, represented in terms of attributes, and a state-based
J. Manner, S. Haarmann, S. Kolb, N. Herzberg, O. Kopp (Eds.): 13th ZEUS Workshop,
ZEUS 2021, Bamberg, held virtually due to Covid-19 pandemic, Germany, 25-26 February 2021,
published at http://ceur-ws.org
Copyright © 2020 for this paper by its authors. Use permitted under Creative Commons License
Attribution 4.0 International (CC BY 4.0).
30 Marius Breitmayer et al.
process model describing object behavior in terms of an object lifecycle model.
In a practical application of the PHILharmonicFlows system, which resulted in a
data- and process-aware e-learning system, examples of business objects include
Exercise, Submission, and Lecture [3]. At runtime, when necessary data (i.e., ob-
ject attributes), such as Points or Feedback, become available, this enables the
transition between the states of a lifecycle. Finally, interactions between object
lifecycles are managed by coordination processes [13].
Section 2 discusses related work. Section 3 sketches three approaches for cal-
culating process progress, followed by a description of the monitoring dashboard
in Section 4. Finally, Section 5 provides a summary and outlook.
2 Related Work
In the BPM lifecycle, process monitoring is a key component for analyzing
and improving processes [8,17]. A wide range of approaches and tools exist for
activity-centric processes, especially in the field of process mining [1]. To iden-
tify problems of a process, conformance checking verifies whether a process is
executed as intended by replaying or aligning an event log with the correspond-
ing process model [6]. However, there only exist few approaches for monitoring
data-centric processes, which focus on artifact-based processes [11]. Process mon-
itoring and conformance checking for activity-centric processes mainly focus on
the identification of bottlenecks, deviations between event log and process model,
and the prediction of the next activities [9]. By contrast, our work deals with the
monitoring of object-aware processes and the large process structures emerging
in this context during runtime.
3 Towards Progress Calculation
To understand the calculation of progress metrics displayed in the dashboard (c.f.
Fig. 1), this section describes the three approaches implemented for calculating
the progress of object lifecycle instances. Though none of the described calcula-
tions covers all aspects, their combination (e.g., individually weighted average)
constitutes a good approximation and is therefore supported by the dashboard.
Approach 1 (Lifecycles). This approach divides the number of completed
lifecycle steps by the number of total lifecycle steps for each object instance. Note
that this calculation is non-trivial as different steps may be executed depending
on data (and routing decisions). Therefore, the number of completed steps may
vary depending on the path taken through the lifecycle. To reconstruct the latter,
for each lifecycle step it is checked whether a value is assigned in the event log.
This allows determining the actual number of steps executed for each object
instance, which then may be divided by the number of remaining steps.
Approach 2 (Relations). Using the hierarchies specified in the data model,
the progress of an object instance can be calculated as the average progress of
its lower-level instances. This approximates how constraints between states of
object instances influence the progress of the overall object-aware process.
A Dashboard-based Approach for Monitoring Object-Aware Processes 31
Approach 3 (Temporal Distance). This approach calculates the temporal
distance between object instantiation and a certain point in time in the event
log. To calculate the progress, this distance may then be divided by the average
duration of all completed object instances of the same type.
4 Monitoring Dashboard
The monitoring of an object-aware process needs to consider the information
contained in its data model (e.g., objects, attributes, relations, hierarchies), its
object lifecycles (e.g., steps and states), and its coordination process [10] in
conjunction with the data recorded during process execution (e.g., event logs).
We created a monitoring dashboard that allows us to display various aspects of
an object-aware process (see Fig. 1). As input, this dashboard takes an object-
aware process (i.e., the data model, lifecycles and coordination process) and
the event log created during process enactment. Note that event logs of object-
aware processes differ from the ones of activity-centric processes. While the latter
consists of case-related activity events enriched with additional information [1],
a data-centric event log contains information about object instances (e.g., ids,
object types, attributes and their values), lifecycle processes (e.g., steps, states,
state changes), and object interactions. Using this information, the monitoring
dashboard can provide an overview of the overall object-aware process (see Fig.
1), as well as drill-down and roll-up functions for inspecting selected aspects.
The dashboard was evaluated for two processes with corresponding event logs.
Fig. 1. PHILharmonicFlows Monitoring Dashboard
32 Marius Breitmayer et al.
1. The tree navigation element lists the various object types (e.g., Lecture,
Exercise, Submission) derived from the data model, together with the corre-
sponding instances (e.g., “Databases” and “Information Systems” as instances
of object Lecture) recorded in the event log. Additionally, the tree navigation
displays lower-level instances of any object instance that can be derived from
the hierarchical structuring of the objects in the data model (e.g., “Exercise1”
as lower-level instance of the “Database” lecture).
2. The sunburst chart provides an overview of selected objects, together with
their instances and corresponding lower-level instances. Each layer contains
the object instances belonging to the same hierarchical level (e.g., all ex-
ercises of a lecture series). However, when the number of object instances
grows, the ratio of each individual instance shrinks. To tackle this issue, we
implemented “drill-down” and “roll-up” functions. This allows inspecting in-
dividual process aspects (e.g., object instances) instead of the entire process.
Additionally, color coding of elements enables fast bottleneck identification.
3. The bar chart provides a quick and easy method to either compare the
progress of similar object instances (see Fig. 1) or the average progress of
instances of the same type. We added “drill-down” and “roll-up” functions.
4. The time slider allows displaying the state of the object-aware process at any
point in time based on recorded event logs. If a point in time other than the
most recent one is selected, the status of displayed instances is reconstructed
through partial event log replay. As a result, all dashboard elements display
the state of the process at the selected point in time. This allows replaying
the process as well as detecting former issues that have been resolved.
5. The list of object instances displays additional information of the object in-
stances being of interest. The table may be sorted according to any criterion.
6. The anomalies table lists object instances that are likely to be outliers. To
identify these outliers, the dashboard allows for the comparison with a rel-
ative or absolute threshold. For relative comparison, this is accomplished
using the following formula:
(1 + threshold(%)
100 ) < InstanceV alue InstanceV alue
ObjectAverage or ObjectAverage < (1 −
threshold(%)
100 )
For absolute comparison, the difference between an object instance and the
object median is calculated and compared with the absolute threshold (e.g.,
3 days). To enable a more sophisticated outlier detection, the dashboard is
able to combine both methods using AND/OR operations.
5 Conclusion and Outlook
This paper presents a monitoring dashboard for object-aware processes which
enables advanced monitoring functions. It combines knowledge from the object-
aware process model with information about the execution to visualize the states
of all object instances at any moment during the process execution. This allows
detecting bottlenecks and outliers. Additionally, the dashboard was tested with
two real-world processes and corresponding event logs. In future work, we will
improve outlier detection based on more sophisticated algorithms and incorpo-
rate different user perspectives (i.e., personalized monitoring views).
A Dashboard-based Approach for Monitoring Object-Aware Processes 33
Acknowledgments This work is part of the ZAFH Intralogistik, funded by the
European Regional Development Fund and the Ministry of Science, Research and
Arts of Baden-Württemberg, Germany (F.No. 32-7545.24-17/12/1)
References
1. van der Aalst, W.M.P.: Process Mining: Data Science in Action. Springer (2016)
2. van der Aalst, W.M.P., Weske, M., Grünbauer, D.: Case handling: a new paradigm
for business process support. DKE 53(2), 129–162 (2005)
3. Andrews, K., Steinau, S., Reichert, M.: Engineering a highly scalable object-
aware process management engine using distributed microservices. In: Int’l Conf
on CoopIS’18. pp. 80–97 (2018)
4. Andrews, K., Steinau, S., Reichert, M.: Enabling runtime flexibility in data-centric
and data-driven process execution engines. Information Systems p. 101447 (2019)
5. Bülow, S., Backmann, M., Herzberg, N., Hille, T., Meyer, A., Ulm, B., Wong,
T.Y., Weske, M.: Monitoring of business processes with complex event processing.
In: Business Process Management Workshops. Springer (2014)
6. Carmona, J., van Dongen, B., Solti, A., Weidlich, M.: Conformance Checking.
Springer (2018)
7. Cohn, D., Hull, R.: Business artifacts: A data-centric approach to modeling busi-
ness operations and processes. IEEE Data Eng. Bull. 32(3), 3–9 (2009)
8. Dumas, M., Rosa, M.L., Mendling, J., Reijers, H.A.: Fundamentals of Business
Process Management. Springer, 2nd edn. (2018)
9. Jorbina, K., et al: Nirdizati: A web-based tool for predictive process monitoring. In:
BPM Demo Track and BPM Dissertation Award (CEUR Workshop Proceedings,
Volume 1920), pp. 1–5 (2017)
10. Künzle, V., Reichert, M.: PHILharmonicFlows: towards a framework for object-
aware process management. J of Soft Maint & Evo 23(4), 205–244 (2011)
11. Meroni, G.: Artifact-driven Business Process Monitoring. Ph.D. thesis, Politecnico
di Milano Milan Italy (2018)
12. Reichert, M., Weber, B.: Enabling Flexibility in Process-Aware Information Sys-
tems: Challenges, Methods, Technologies. Springer, Berlin-Heidelberg (2012)
13. Steinau, S., Andrews, K., Reichert, M.: Modeling process interactions with coor-
dination processes. In: CoopIS’18. pp. 21–39. LNCS, Springer (2018)
14. Steinau, S., Andrews, K., Reichert, M.: The relational process structure. In: CAiSE
2018. pp. 53–67. No. 10816 in LNCS, Springer (2018)
15. Steinau, S., Andrews, K., Reichert, M.: Executing lifecycle processes in object-
aware process management. In: Data-Driven Process Discovery and Analysis. pp.
25–44. Springer (2019)
16. Steinau, S., Marrella, A., Andrews, K., Leotta, F., Mecella, M., Reichert, M.:
DALEC: A framework for the systematic evaluation of data-centric approaches
to process management software. Softw & Sys Modeling 18(4), 2679–2716 (2019)
17. Weske, M.: Business Process Management: Concepts, Languages, Architectures.
Springer (2019)