|This content applies solely to Process Mining, which must be purchased separately from the Appian base platform.
This page describes how to use record event data to mine Appian applications.
You can mine data directly from your Appian applications when you use Records as a data source. This allows you to seamlessly connect to your application data and leverage record events to more easily mine your business processes.
Record events allow you to track how different parts of your business process operate. As a best practice, each record type in your application should have record events configured. This means that each of your major business concepts should have a corresponding Event History record type and Event Type Lookup record type. When you generate these record types as part of your event configuration, they are specifically structured so you can easily prepare the data for mining—allowing you to gather insights faster.
The generated Event History and Event Type Lookup record types provide your core event data, like who performed an activity and when. If there are other record types that include attributes that you want to explore, like order status, shipping date, or total cost, you can add data from those record types as well.
If you are not using record events, you still have the flexibility to use record types in process mining. Before using those record types, review What makes a good data set? to ensure you can easily mine that data.
Appian version 23.2 or later to mine data using record events.
If you are not using record events, you can import record data beginning in Appian version 22.3.
It's easy to mine data from an Appian application when you use the Records data source. This allows you to import record data into data sets and transform that data to meet your mining needs. When you import data from a generated Event History record type and Event Type Lookup record type, it's even easier to prepare your data since the data structure is specifically created for mining. Once the data is transformed, you'll execute the transformation project to transform your record data into an event log and load it into Process Mining.
The following instructions detail how to import and transform data from a generated Event History record type and Event Type Lookup record type.
Add a Records data source to import record data into Mining Prep.
Before you begin these steps, make sure you have a service account API key.
To add a Records data source:
Configure the following properties:
|Enter the name you want to display in the list of data sources.
|Enter the URL of your Appian Designer environment. For example,
|Enter the service account API key.
Once you've added the Records data source, you can import the record data as data sets. You can only import data from record types that your Process Mining service account can access. You can import data from multiple record types, and each record type will import into a unique data set.
If you configured record events, you should import both the Event History record type and the Event Type Lookup record type for the processes you’re investigating. For example, if you want to investigate your order-to-cash process, you could import the Order Event History record type and the Order Event Type record type.
To import data from your event record types:
Consider importing any additional record types that can enrich your data set. For example, you may consider importing the record type that you’re tracking events for, like Orders, Customers, or Order Items.
Record types that contain close to 2 million rows or more than 50 columns may take over an hour to import.
Once you've imported the record data into different data sets, you can create a transformation project to transform the data sets into an event log for mining.
To create a transformation project:
Now that you've created a transformation project, you'll select the data to transform and apply transformation actions.
If you imported an Event History record type and an Event Type Lookup record type, you’ll need to combine the data into a single data set to see all event information in one place. You’ll also need to identify the data properties, like the case ID, events, and any case attributes.
If you are not transforming data from the generated Event History record type and Event Type Lookup record type, see Transform and Load Data for general instructions on how to transform data.
First you need to populate your project with a data set. You can only populate this building block with one data set to start, but you can add data from other data sets later.
As a best practice, start by using the Event History data set as your data source.
To select the Event History data set as your data source:
Now that you've selected the Event History data set as the data source, you'll need to add the Event Type Lookup data set to the project so you can display the different activities that occur in the events. Adding this data will combine the two data sets into one.
To combine the Event History and Event Type Lookup data sets:
Configure the following properties:
|Select key column from upper table
|Select the eventTypeId field.
|Select matching key column from lower table
|Select the Id field.
Add any additional data to the project that will enrich your analysis.
Next you'll need to identify the case ID, which is the unique identifier of the business case.
If you imported the generated Event History record type, this will be the recordId field. This is the field that links each event to a specific record in Appian.
To identify the the case ID in the Event History and Event Type Lookup data sets:
Identifying the events is one of the most important parts of mining prep. This specifies the fields that store the types of activities that can occur, the timestamp when the event occurred, and who started it.
To identify the events in the Event History and Event Type Lookup data sets:
Order activities. This is only used in Mining Prep and doesn't display in Process Mining.
For Categorical Attribute Columns, select the user field from the Event History data set.
All Event Attributes are optional, but recommended. Event attributes allow you to specify any additional characteristics related to an activity.
If you imported other record types, like the record type for which you're tracking events, you can add case attributes to the project to further enrich your data. While event attributes describe characteristics of an activity, case attributes describe characteristics of a case.
Each case can have one or more attributes that describe it in more detail. For example, in an transformation project containing data about your order-to-cash process, you may want to include order name, customer name, or order cost as case attributes.
For more information, see Case Attributes.
Once you've finished setting up your transformation project, you'll execute the project to apply the transformation steps and load the event log into Process Mining.
To load and execute the transformation project:
That's it! Now you can navigate to your Process Mining environment to start investigating your process.
To ensure you always have the latest data from your record types, we'll automatically update the data in your records data sets whenever you execute or schedule updates to a transformation project.
You can manually reload your data for in-progress transformation projects. Since you typically won't set up a transform and load schedule until your project is complete, manually reloading your data keeps it fresh during this stage.
To manually update your data set:
Mine Appian Applications