Data preparation

To find solutions to your business challenges, it is crucially important to identify the required data correctly. There are two common approaches to process data identification:

  • Process-oriented data identification

    Define the beginning and end of your process and all tasks and events in between.

  • Business object-oriented data identification

    Identify business documents that are involved in the process - for examples orders or invoices - and follow the lifecycle of these documents.

When you have defined the scope of the data with one of the approaches, you can locate it by its IT footprint in the corresponding databases of your application landscape.

Transforming your data

After you have identified the data, you need to transform it to fit the structure Signavio Process Intelligence requires. The system requires two dimensions (data logs), which serve different purposes and should therefore be extracted separately.

  • Event log

    The event log lists the core properties of all events that occurred as part of the specified process.

  • Case attribute log

    The case attribute log lists the core properties each case (instance) of the specified process. These properties, like for example case creator, apply to a case in general, and are not dependent on specific events.

Data requirements

Below you will find a precise specification of how you should provide the event and case attribute logs so that Process Intelligence can import the data.

The event log needs to contain the following properties for each event it contains:

  • CaseId: The unique identifier of the case the event is assigned to.
  • EventName: The name of the event.
  • Timestamp: The exact date and time when the event occurred, in the format YYYY-MM-DDTHH:MM:SS.

The following table contains an example excerpt of an event log:

Event log
CaseId EventName Timestamp
100430031000112060012015 Create FI invoice by vendor 2014-11-20T00:00:00.000
100430031000112060012015 Post invoice in FI 2014-11-20T00:00:00.000
100430031000112060012015 Clear open item 2015-01-12T23:59:59.000

For each case, the case attribute log must contain the caseId property. It can list additional properties that provide more insight into the cases of the process. For each property with the exception of caseId, you need to specify a data type. The following data types are available:

  • CHOICE, for any text value, including multiple choice, such as country names or language.
  • DOUBLE_PRECISION (number), for example 12.34
  • TIMESTAMP_WITH_TIME_ZONE, for example 2018-01-01T12:45:22. (Note that the name is misleading. The structure of the time stamp needs to be YYYY-MM-DDTHH:MM:SS (without time zone)),
  • CURRENCY, for example 5.21.
  • DURATION (in milliseconds), for example 268488000.

The following table contains an example excerpt of a case attribute log:

Case attribute log
CaseId Invoice due date (TIMESTAMP_WITH_TIME_ZONE) Invoice value (CURRENCY) Material group (CHOICE)
100430031000112060012015 2015-11-06T14:32:37.000 175.38 A001
100430031000112070012015 2015-06-10T06:45:38.000 0.41 -
100430031000112080012015 2015-06-19T14:48:50.000 0.20 -

You need to provide the data as two separate CSV files. The CSV’s value separator must be a semicolon (;).

Below you see an excerpt of an event log:

CaseId;EventName;Timestamp
100430031000112060012015;Create FI invoice by vendor;2014-11-20T00:00:00.000
100430031000112060012015;Post invoice in FI;2015-01-08T14:26:02.000
100430031000112060012015;Clear open item;2015-01-12T23:59:59.000

This is an excerpt of a case attribute log:

CaseId;invoice duedate (TIMESTAMP_WITH_TIME_ZONE);invoice value (CURRENCY);material group (CHOICE)
100430031000112060012015;2015-11-06T14:32:37.000;175.38;A001
100430031000112070012015;2015-06-10T06:45:38.000;0.41;
100430031000112080012015;2015-06-19T14:48:50.000;0.20;

Note that the data types are provided in parentheses after the property name.

Uploading data

Once your data is prepared, open Process Intelligence.

The next step depends on if you have existing processes:

  • If you have not previously created a process to investigate, click the blue Upload new process data button.
  • If you already have existing processes, but want to upload data for a new process, click the white Create new process button on the right-hand side.
  • If you want to upload new data for an existing process, select the process from your process list and click Process data settings at the bottom of the navigation pane.

If you’re starting a new process, you will be prompted to add a title. We recommend using the question that you are investigating (i.e. “Why does our London office have slower response times than our Berlin office?”).

Afterwards, you will be brought to the “Manage data” page. Here, you can upload your prepared event log and attribute log (if using), as well as set any custom case attributes. Select the files you want to upload. Process Intelligence will preprocess your data automatically as part of the uploading process.

You can check the status of your file in the Status column. If the upload fails, a red X will appear.

../_images/upload_failed.png

Click on it to get more information about what went wrong.

../_images/upload_failed_message.png

Deleting process data

To delete process data, click Manage process data in the navigation bar on the left-hand side. From there, click the trash can icon next to the name of the file you would like to delete.

If you want to delete all of your process data at once, click the Delete process data button on the bottom left-hand side of the screen.

A confirmation dialog will appear.

../_images/confirmation.png

Note

Deleted data is not archived. If you delete process data and decide you need it again at some point in the future, you’ll need to re-upload it.