ion perspective, we can realize that all the points of view share a common ground: the business process model and its instantiation are in the kernel of all of them. In this paper, we propose the use of a Business Process Instance Metamodel, which serves as a common interface to make independent the applications producing business process data from those applications that consume and exploit it. A tool has been implemented as a proof of concept to facilitate the matching between data from di erent data sources and the metamodel.
Activities undertaken by companies can be choreographed by business processes to achieve their goals. Many companies use Business Process Management Systems (BPMSs) to support the automation of business processes, but others use software applications that are implemented ad hoc, having their own data sources with a domain-speci c structure. As a consequence, business data cannot be exploited in a general way and, the analysis carried out using the stored data must also be developed ad hoc.
Some of the most common ways of exploiting data generated during process
execution are: the creation of execution traces for process discovery algorithms to
obtain the system behavior [
On the one hand, the need for a data representation model that let
applications that generate business data be independent from those that consume them
was detected in [
In this paper, we propose the use of a Business Process Instance Metamodel
as an intermediate layer to bring closer the domain speci c data produced by
business processes and how they can be exploited. The de nition of an
intermediate metamodel helps us to abstract domain speci c knowledge from the
analyzed data, being easier to apply di erent process analytics techniques. The
approach is based on the de nition of mappings between the data source and the
expected elements in a Business Process de ned as a Business Process Instance
Metamodel. The bene ts produced by the use of an intermediate metamodel are
the reduction of the analysis time, and the exploitation of data in a more
appropriated way [
The paper is organized as follows: Firstly, Section 2 introduces the main ideas of our approach and describes brie y the Business Instance Metamodel. Secondly, Section 3 presents a tool implemented as a proof of concept to de ne the matching between an Oracle database and the Process Instance metamodel. And nally, the paper is concluded and some further work is pointed out. 2
Business process data exploitation depends highly on the technology that supports business process execution as well as the way data is structured. Some companies have BPMSs to support the execution, but others have software applications that have been implemented ad hoc. As a consequence, there is no standard approach to exploit these data. Thus, for example, the data conversion needed to generate an event log for a process mining tool is totally di erent if data are stored in a relational database or the data storage of a SAP system.
To make data exploitation technology-agnostic, our approach is inspired by the guidelines provided by Model-Driven Architecture to structure speci cations, in such a way that we propose a Business Process Instance Metamodel that helps us to separate the technological details and the structure of the information from the data itself. In other words, the business process instance metamodel can be seen as an intermediate artifact that is not dependent on the business domain nor focused on a subset of tools. Thus, it allows us to make applications that produce business process data independent from those applications that consume them. Figure 1 depicts how the process instance metamodel acts as an interface for both, data producers and consumers.
The Business Process Instance Metamodel depicted in the center of Figure 1
is detailed in Figure 2. The metamodel has been speci ed with EMF [
The root of the metamodel is the ProcessEngine metaclass and represents the BPMS or software application that is in charge of processes execution. The process engine can be in charge of di erent processes. The ProcessDe nition metaclass represents the formal de nition of the process, that is, what we call the Business Process Model. One business process can be executed many times and the ProcessInstance metaclass models these executions or instances. A business process is composed of di erent activities and the Activity metaclass models them. Finally, the ActivityInstance metaclass represents the execution of an activity and it is related to the Activity metaclass (note that an activity may be executed many times) and to the ProcessInstance metaclass (an activity may be executed in the context of di erent business processes). This metamodel allows us to exploit business data in di erent contexts independently of the storage technology and structure. We only need to de ne mappings from the concrete technology to the Process Instance Metamodel. Then, the information stored as a BP instance Metamodel may be used to generate event log traces (both in XES or MXML format), to be queried for decision-making or to be semantized and to reason about it. 3
We have implemented a tool that assists business experts in de ning the mapping between the data repository and the Process Instance Metamodel, facilitating the later data analysis. The tool allows, once the data repository and metamodel are connected by the mapping de nition, to analyze and use the data from a Business Process point of view and exploit those data by applying any of the techniques used in the di erent business analytics contexts. The tool has been developed as a web application. Figure 3 shows a screen-shot that captures the mapping de nition process. Further details about the tool can be found in http://www.idea.us.es/portfolio-item/process-data-matching-tool/.
This paper shows how the use of an intermediate metamodel can help to standardize the exploitation of business process data by de ning a common infrastructure that may be used in di erent business process analytics contexts.
As further work, we consider interesting to enrich the way of de ning the matching, making more exible the tool and allowing the building of more complex processes and exploitation of more complex data sources. This work has been partially funded by the Ministry of Science and Technology of Spain (TIN2015-63502-C3-2-R and TIN2016-75394-R) and the European Regional Development Fund (ERDF/FEDER).