=Paper=
{{Paper
|id=Vol-3783/paper_359
|storemode=property
|title=3DCR: A Tool for Immersive Process Mining
|pdfUrl=https://ceur-ws.org/Vol-3783/paper_359.pdf
|volume=Vol-3783
|authors=Juanita Caballero-Villalobos,Simon James Jensen,Hugo-Andrés López-Acosta
|dblpUrl=https://dblp.org/rec/conf/icpm/Caballero-Villalobos24a
}}
==3DCR: A Tool for Immersive Process Mining==
https://ceur-ws.org/Vol-3783/paper_359.pdf
3DCR: A Tool for Immersive Process Mining
Juanita Caballero-Villalobos1 , Simon James Jensen1 and Hugo A. López1
1
Technical University of Denmark, DTU Compute, Denmark
Abstract
3DCR is a tool to enhance the comprehension and engagement of declarative process models
via 3D gaming environments. Integrating with an industrial declarative process modeling
language (DCR graphs), our tool enables a tailored and adaptable environment where novel
representations of declarative process elements (e.g., events and causal relations) can be
explored, modified, and enriched with domain-specific representations. 3DCR helps in the
elicitation of process behaviors, the discovery of processes from user interactions, and the
simulation/upskilling of stakeholders in new process variants. Initial usability experiments
indicate positive results regarding usefulness, engagement, and learnability compared to 2D
process model representations. In this paper, we describe the architecture, functionality, and
maturity of 3DCR.
Keywords
Process Mining, Declarative Process Models, Virtual Reality (VR), Human-Centered BPM,
DCR graphs
Metadata description Value
Tool name 3DCR Beta
Current version 1.0
Legal code license MIT License
Languages, tools and services used C#, ShaderLab, HLSL, Unity Engine 2022.3.13f1
Supported operating environment Microsoft Windows, Linux & Mac
Download/Demo URL https://bit.ly/demo3DCRBeta
Documentation URL https://bit.ly/3DCRdocs
Source code repository https://bit.ly/sourcecode3DCRBeta
Screencast video https://bit.ly/3DCRvideos
1. Introduction
Process mining has traditionally relied on 2D representations to visualize and analyze
process models, with examples such as Directly Follows Graphs, Petri Nets, or BPMN
models. While 2D representations are effective in many scenarios, their adaptation
to declarative settings, characterized by causal relations, may not be straightforward.
Indeed, classic graph-theoretic representations may be semantically perverse, meaning
that the notation conveys a message different than the semantic model of the declarative
process language [1, 2, 3]. Moreover, the complexity of the relations, dynamic constraints
ICPM 2024 Tool Demonstration Track, October 14-18, 2024, Kongens Lyngby, Denmark
$ jcavi@dtu.dk (J. Caballero-Villalobos); hulo@dtu.dk (H. A. López)
0000-0002-4915-0961 (J. Caballero-Villalobos); 0000-0001-5162-7936 (H. A. López)
© 2024 Copyright for this paper by its authors. Use permitted under Creative Commons License Attribution 4.0
International (CC BY 4.0).
CEUR
ceur-ws.org
Workshop ISSN 1613-0073
Proceedings
handling, and event dependencies can become difficult to analyze with the current tools
[4, 5]. These challenges underscore the need for more advanced visualization techniques
that might enhance user comprehension and engagement.
Existing technologies have provided interactive representations of process models
using Virtual reality (VR) and augmented reality (AR) environments to make complex
data more accessible [4, 5]. However, these technologies face cost-related, technical-
related, user-related, and data-related challenges, highlighting the lack of domain-specific
customizations as a crucial factor for enhancing the adoption and benefits of AR and VR
in process support [6]. Therefore, a balanced approach that leverages the strengths of
interactive technologies while addressing their limitations is needed.
We presented 3DCR, a gaming-based tool that offers a more intuitive and engaging
way to represent and execute declarative process models based on domain-specific repre-
sentations. The virtual environment incorporates visual and sound aids and interactive
elements, making it easier for users to understand complex process dynamics. The tool
contributes to two use cases in process mining: elicitation and discovery of process
variants using unrestricted process models and simulations of process models.
Digital traces of the simulations (logs) can be analyzed and extracted in real-time
analysis with SQL in Unity Analytics Dashboard and exported as CSV files for process
discovery and conformance checking. Furthermore, the 3DCR environment could be
adapted to trace simulation and digital twin representation by adding more variables
(e.g., time, data, locations, novel rules) and Computer-Aided Design (CAD) models to
emulate the process environments and impact variables more accurately. To sum up,
the tool offers an opportunity to strengthen the portfolio of applicable process mining
techniques by integrating VR environments.
2. Architecture
3DCR tool enhances user engagement and comprehension of declarative process models
(e.g., Dynamic Condition Response Graphs [7]) through 3D visualizations (e.g., Unity
Games). The architecture of the tool (Fig.1) was designed around an extended Model-
View-Controller (MVC) pattern, integrating both a process engine and a 3D engine to
facilitate immersive process simulation.
The business process model data is processed by the Process Engine:, which manages
the execution and state of the declarative process model, ensuring that the process logic
is correctly followed and updated based on user interactions and process rules.
The processed data feeds the Application Controller, which initializes the environment
and handles synchronization between the game view and the process model. as a central
hub that coordinates the interactions between the user, the process model, and the 3D
representations (activities, visual effects, camera views) that are rendered by the 3D
engine. Moreover, the 3D Engine also updates the 3D views based on user interactions
and the current state of the process model. In addition, the Activity Engine receives the
same data to render activities in the 3D view, selecting domain-specific representations
according to the available prefabs. Furthermore, it integrates a standard (canonical) view
Figure 1: Application Architecture
of activities with their markings, adapting the visual representation of activities based on
their state (e.g., included, excluded, pending). Furthermore, the 3D View comprises both
outputs (3D Engine and Activity Engine), presenting the game’s user interface, including
activities, visual effects, and camera views. Finally, user interactions are automatically
collected and can be analyzed directly with SQL features of Unity Cloud or exported in
CSV format for further analysis with process mining tools.
3. Functionality
The setup process starts with the model of the process in the Dynamic Condition Response
Graphs (DCR) Simulator [8]; for our example, we chose the sickness registration process
(Fig. 2). Then we export its logic in XML format and upload them in 3DCR Game Data
Folder to update the initial activities representation (cell maps contents) (Fig.2 ) and
the process model according to the defined rules, data, and roles [7]. Users can modify
the 3D Engine and Activity Engine if they want additional customization.
Through 3D gaming environments for business process model representations and
executions, users can explore the DCR activities rendered and interpret straightforwardly
the relationships guided by the visual aids included in the tool visual effects aid (Fig.3.)
Furthermore, the tool aims to increase user engagement and learnability through
interactive elements and domain-specific 3D representations. Users can choose the specific
domain representations using the main menu (Fig.4). Overall, the 3D representations
added specialization for depicting activities and the potential to show causal relations
through animations, but this complexity could increase semantic ambiguity [9].
Nevertheless, to enhance process granularity understanding, 3DCR provides multiple
perspectives views, including first-person, third-person, and global views. Users can
switch between different views to gain insights into the process model, such as a detailed
view of specific activities or a broader overview of the entire process (Fig.5).
After the 3D Environment setup, users can interact with the process models, analyze
Figure 2: 2D (left) vs 3D (right) representations of the sickness registration process.
Figure 3: Marking domain-specific representations. From left to right, (1) Enabled activity, (2)
Disabled activity, (3) Excluded, not executed, not pending, (4) Included, executed, not pending, (5)
Included, not executed, pending.
Figure 4: 3DCR Generator Beta - Domain representations configurator and exemplary agents.
the context variants, execute actions, and observe their immediate effects within the 3D
environment. Indeed, simulating different scenarios by altering process parameters in
the XML file and 3D Engine (e.g., rules, roles, activities) and analyzing the outcomes
contribute to identifying potential bottlenecks and inefficiencies, enabling more informed
decision-making. Moreover, 3DCR includes the Unity Analytics plugging, which auto-
matically gathers logs of user interactions and properties when someone interacts with
their game. The dashboards can be accessed by the organization owner (unity developer)
from the online cloud Unity Website; the logs allow one to monitor specific in-game user
actions by defining custom events, due to the tool allowing one define more events to
capture additional logs of enriching the existing logs with additional parameters (e.g.,
add contextual information, counters, flags). In the online version, SQL queries can
be executed to automatically explore the logs and export them in CSV format (Fig.6).
(a) First Person View (b) Global perspective (c) Third person view
Figure 5: 3DCR Views
Consequently, the collected data could also be processed for external process mining
tools. Fig. 7 and 8 show the integration with Disco.
(a) SQL (b) Custom Events (c) Custom Parame-
ters
Figure 6: 3DCR Beta - Unity Analytics
Figure 7: Excerpt DISCO process variant Figure 8: List of events - Case Example
4. Maturity
3DCR was developed in Unity using the editor version 2022.3.13f1. We evaluated the
maturity by verifying basic functionality, such as successful deployment in Play Unity,
launch, and responsive user interface as well as usability was validated in [9] with 8
participants (4 workers, 3 DCR novice users, 1 DCR expert). Usability metrics indicate
that while 70% of tasks were completed without assistance, there was a 20% error rate,
often due to confusion around UI elements and the sequence of activities. Additionally,
40% of users found the naming conventions confusing, and 30% expressed uncertainty
about the next steps when deviating from the intended process. Improving the clarity
of feedback and guidance could potentially reduce confusion and errors. Moreover,
perspective switching (views) added a 15-20% time burden for some users. The learning
curve was significant, with inexperienced users taking 30-35% longer to complete tasks
than those familiar with similar environments.
Furthermore, we evaluated the performance of the main scene and the menu scene Fig.4
after 30 continuous simulations on a system running (Windows 11 Enterprise, version
23H2, 64-bit, 13th Gen Intel(R) Core(TM) i7-1365U 1.80 GHz, and 32,0 GB RAM ),
including an analysis of memory allocation, heap utilization, Unity object categories, and
CPU usage. Key findings of the main scene show a total allocated memory of 2.85 GB,
with 2.20 GB being resident on the device, distributed across various memory categories;
managed Heap Utilization of 186.8 MB, with Objects consuming the most significant
share; breakdown of Unity object categories shows RenderTexture, Shader memory, and
Texture2D as the largest consumers; and a CPU usage metrics indicate the application’s
ability to maintain real-time performance at 60 FPS, 30 FPS, and 15 FPS, with rendering
statistics and frame times suggesting potential areas for optimization.
5. Conclusions and Future Work
We proposed enhancing the 3DCR as a tool for simulation, elicitation, and education of
declarative process models via virtual and immersive environments. In future work, we
would like to explore the combination of multiple dimensions and the effect of their 3D
representations. These include time, locations, layout, simulation of multiple instances,
and multiple interacting agents and process models [10, 11]. Moreover, we would like to
expand the behaviour of the simulator to support data input parameters [12]. Finally,
incorporating advanced VR and AR technologies could increase the immersiveness of
the tool, creating a more engaging and interactive user experience with the potential
for voice commands or gesture-based interactions, making the tool more intuitive and
accessible to a broader range of users.
Acknowledgments
This work was supported by the research grant “Center for Digital CompliancE (DICE)”
(VIL57420) from VILLUM FONDEN.
References
[1] K. Figl, C. Di Ciccio, H. A. Reijers, Do declarative process models help to reduce
cognitive biases related to business rules?, in: ER, Springer, 2020, pp. 119–133.
[2] H. A. López, V. D. Simon, How to (re) design declarative process notations? a view
from the lens of cognitive effectiveness frameworks, in: POEM, CEUR-WS, 2022,
pp. 81–97.
[3] D. M. T. Trinh, A. Abbad-Andaloussi, H. A. López, On the semantic transparency
of declarative process models: The case of constraints, in: International Conference
on Cooperative Information Systems, Springer, 2023, pp. 217–236.
[4] R. Oberhauser, C. Pogolski, A. Matic, VR-BPMN: Visualizing BPMN Models in
Virtual Reality, Lecture Notes in Business Information Processing 319 (2018) 83–97.
[5] G. Grambow, D. Hieber, R. Oberhauser, C. Pogolski, A Context and Augmented
Reality BPMN and BPMS Extension for Industrial Internet of Things Processes, in:
BPM Workshops, Springer International Publishing, Cham, 2022, pp. 379–390.
[6] A. Mostafa, A. Alaqeeli, Benefits and challenges of integrating iot, vr & ar in the
bim-based facility management process: Literature and case-based analysis, Journal
of Engineering Research 6 (2022) 25–40.
[7] T. T. Hildebrandt, R. R. Mukkamala, Declarative event-based workflow as dis-
tributed dynamic condition response graphs, in: PLACES, volume 69 of EPTCS,
2010, pp. 59–73.
[8] M. Marquard, M. Shahzad, T. Slaats, Web-based modelling and collaborative
simulation of declarative processes, in: Business Process Management, Springer,
2015, pp. 209–225.
[9] S. J. Jensen, H. A. López, Towards immersive environments for declarative process
models, in: International Conference on Business Process Management, Springer,
2024, pp. 214–231.
[10] T. Hildebrandt, H. A. López, T. Slaats, Declarative choreographies with time and
data, in: BPM, volume 490, Springer, 2023, pp. 73—-89.
[11] S. Debois, H. A. López, T. Slaats, A. A. Andaloussi, T. T. Hildebrandt, Chain of
events: modular process models for the law, in: iFM, Springer, 2020, pp. 368–386.
[12] R. Strømsted, H. A. López, S. Debois, M. Marquard, Dynamic evaluation forms
using declarative modeling, in: BPM, CEUR Workshop Proceedings, 2018, pp.
172–179.