Skip to content

Latest commit

 

History

History
112 lines (66 loc) · 5.27 KB

process-mining-transform.md

File metadata and controls

112 lines (66 loc) · 5.27 KB
Raw
title description author contributors ms.subservice ms.topic ms.date ms.author ms.reviewer search.audienceType
Transform and map data
Learn how to transform and map data with the process mining capability in Power Automate.
donums
donums
v-aangie
process-advisor
conceptual
07/18/2023
derahonuorah
angieandrews
flowmaker
enduser

Transform and map data

After you select the data source you want to use, you're taken to the Power Query Editor. The Query Editor is a powerful tool to transform your data. To learn more, go to The Power Query user interface.

Here are some reasons why you might want to transform the data:

  • You might not be interested in all the activity names that are logged, and so you want to filter for specific activity names that are important to the process you're trying to mine.

  • You might want to rename some of the activity names to be more descriptive and understandable. This is often not the case with names in a database.

  • You might be interested in only a specific date range, and not the entire history of data.

  • You might want to combine multiple ID columns to form the case ID. This is often done when the ID you want to use for process mining doesn't exist or is a combination of multiple IDs in the application. For example, when a customer files support tickets, the support ticket might be assigned to multiple customer service agents. If you want to analyze how each agent handles each ticket, you would combine the agent ID and the ticket ID into the case ID.

Filter activity names

  1. Select the caret next to the activity name column to bring up the sort and filter menu.

  2. If there's a message that says List may be incomplete, select Load more.

  3. Select only the activity names that you want to analyze. Uncheck any name you want to exclude.

    :::image type="content" source="media/process-mining-transform/filter-activity.png" alt-text="Screenshot of filtering activity names.":::

  4. Alternatively, you can use the Text filter menu for more advanced filtering. For more information on filtering by value, go to Filter by values in a column.

Rename activities

  1. Ensure that the Activity Name column is selected.

  2. Above the toolbar, select the Transform tab.

  3. On the toolbar, select Replace values.

  4. Under Values to find, type the activity name as it appears in the data source that you want to replace.

  5. Under Replace with, type the activity name you want to show in the process map.

  6. Repeat this process for all the values you want to replace.

Reduce the number of total records

One strategy for reducing the total number of records is to use only the latest records. To do this, you need to first sort the data by time.

  1. Open the sort and filter menu by selecting the caret next to the startTimestamp column.

    :::image type="content" source="media/process-mining-transform/reduce-records.png" alt-text="Screenshot of reducing the number of total records.":::

  2. To have the most recent records show up first, select Sort descending.

  3. Select OK, and then select Keep rows on the toolbar.

  4. Enter 150000 under Number of rows.

  5. Select OK to filter for the top 150,000 rows.

Combine multiple IDs

You can use Ctrl + click to select multiple columns that you want to combine.

  1. On the Transform tab toolbar, select Merge columns.

  2. (Optional) Select a separator from the dropdown list. Alternatively, you can select none.

  3. Change the name (or make a note of the default name that's generated), so you can select it when mapping to a case ID.

Map data

Mapping tells the process mining capability what columns represent case ID, activity name, or timestamp.

  1. In the toolbar, select Map to entity.

  2. In the left column, if your data only has one timestamp, select the Event log entity. If your data has both a start and an end timestamp, select the Activity log entity.

    :::image type="content" source="media/process-mining-transform/map-cdm.png" alt-text="Screenshot of mapping options.":::

  3. Select each of the dropdown lists under the Query output column heading, and select the corresponding column in your data that maps to:

    1. Activity name (activityName)
    2. Case ID (caseId)
    3. Start timestamp (startTimestamp)
    4. End timestamp (endTimestamp) only if you select Activity log in step 2.

  4. After everything is mapped, select OK to close the map to entity screen.

Select additional columns

When performing process mining, having additional information about a process is especially useful. If your data contains only five or fewer columns outside of the required columns that you've mapped in the previous procedure, nothing else needs to be done. You can select Save and all the columns will be analyzed. You'll also be able to filter on these columns in the Analytics page.

If you have more than five columns:

  1. Select Choose columns in the toolbar.

  2. Deselect any column you don't want to include in the analysis.

    Ensure that the mapped columns (activityName, caseId, startTimestamp, endTimestamp) from the previous procedure are selected. The five additional columns are on top of the mapped columns.