Uploading a Data Set (Manually)
Although it's common to use a connector to move data, let's focus on how to manually upload data.
A new PI process works as your data bucket. It contains the event and the attribute log. Several adjustments to the process setting can be made as well, such as custom attributes, data views, etc.
Continue below to watch a video on how to create a new process.
Step 1 - Create a New Process
Watch the video below on how to create a new process.
Step 2 - Upload the Event Log
Next, look up your event log via the file explorer or just drag and drop your data file into the corresponding area. Afterwards, you can change the format of the timestamp or just select Import.
Step 3 - Upload the Attribute Log
Now we want to upload the attribute log in a similar way. There is also the option to change individual attribute types during the import. If there are no adjustments, accept all suggestions and execute the import.
Step 4 - Create your Investigation
Now your process is ready for the analysis. Back on the list of "all investigations" of this process, we can create a new investigation.
Besides restricting the attribute view, you can also opt to map your process data to an existing BPMN model. This is provided by the SAP Signavio Process Manager.
Now the stage is set for the actual process mining task. Next, we will look at how to use the analytic tools to build up an investigation to target key questions and gain insights.
Starting the Investigation
In PI, an investigation gives insights into the changes needed to improve and optimize your process. An investigation is a direct visualization of the data that is stored in transactional systems.
- Each investigation starts with the Process discovery, which gives insights into the different process variants from the most common path to nonconformance edge cases. It's basically a visualization of all executed events from the event log.
- Chapters lead the investigation with text descriptions of aims, objections, findings, observations etc., and are part of the overall structure of an investigation.
- In Process Mining the term Activity describes the different types of possible events. The term Event is most often used to describe individual executions of activities.
Here is an overview of the most important elements in your investigation.
Select each button to learn about the feature.
Starting the Investigation
Different Widgets for Different Purposes
Different widgets are used to tell stories on your investigation. Their main purpose is to explore and assess the interaction of events and key process factors.
There are different widgets that you can pick from. Choose which widgets you need based on your questions and targets. For example, there are widgets that focus on conformance checks, variance, or performance analyzing.
A widget basically represents a single analytical component.
Some widgets serve to calculate and present special key figures in a simple way, for example, Value or Relate.
Others are especially helpful to illustrate the ratio or distribution of certain performance indicators, for example, Over time, Distribution or Breakdown.
In order to reveal nonconformities or variations in the execution of the process, you can use Process Conformance, Variant Explorer, Process Discovery.
More widgets like Case Table, Activity List or Diagram allow you to extend your analysis with additional explanations or information.
- Processes representing transactional data.
- One investigation covers only one specific topic.
- Chapters breakdown the results for clarity and filtering.
Let's take the first trip through the capabilities of Process Intelligence!