Information Technology Reference
In-Depth Information
logs. We suggest to label such statements as
comments
to recall the data if necessary.
In order to be able to transform results of the qualitative content analysis into process
view logs, post conditions of the analysis are listed in the following.
Post condition 1: Each To-do of each person is considered as a case. A case is the unit of analysis.
Post condition 2: Each case is labeled. The labels reflect the key essence of the To-do. The labels illus-
trate the labels of tasks and activities in the process model.
Post condition 3: A precise coding guideline is elaborated that support comprehensibility and traceabil-
ity.
Post condition 4: Each case is complemented by variables. Variables necessary for process modeling or
log preparation are: Agent (name of the process member), Agent Position (role of the
process member), Tool (tool name), DataInput (data name), DataOutput (data name),
OrganizationOfWork (if delegation name of process member, else (direct work) no en-
try), Connection (person name or role label), Participation (person name or role label),
Decision (String), Recurrence (Boolean), Event (event type), TimeNote (String). Further
variables might be, e.g. Granularity (GranularityLevel), FirstGeneralizationLabel
(String), SecondGeneralizationLabel (String).
Post condition 5: The beginning and the end of the entire process are identified (e.g. from time notes or
from the semantic context).
According to the BPME method, the designer can choose two alternative paths after
analyzing the data. The process view models and the entire process model can be
modeled either (1) manually, (2) using a tool that supports scenario-based process
modeling like GRETA [1], or (3) the designer transforms the captured process view
information into process view logs and entire process logs. In the following we focus
on the latter. We argue that the extraction of the business process model out of
process view logs offers a structured and traceable way of illustrating the identified
process model. We present guidelines for preparing logs based on results of the quali-
tative content analysis of process member To-do's in the next section.
4 Preparation of Logs and Process Mining
We want to know how the processes of the individual process members as well as the
whole process look like. In order to capture all possible variants of performing proc-
ess paths, process view logs for each process participant are prepared for mining. The
preparation of the logs is based on the results of the qualitative content analysis of
To-do lists of process members. We suggest the designer to choose the level of ab-
straction that is included in the original To-do lists (results of the paraphrasing in the
qualitative content analysis), as the To-do's reflect the “mental process steps” of each
process participant. For each To-do list a process view log is elaborated. One output
of the log preparation step are logs that reflect To-do's of a process member in a par-
ticular thematic process (e.g. sale process of a particular organization). The process
view logs reflect process fragment scenarios. The most challenging task for the de-
signer is now to transform reference points (e.g. operations that are connected to a
message event between process participants (send and receive), operations that are
performed by several process participants at the same time, like meetings, and time
notes) into concrete (artificial) start- and end date and time stamps. The procedure of
considering reference points in the process view logs is described in the following.
