Skip to main content
Version: 2.11.X

Pulling in Your Data

In the preceding steps, you've configured Cogynt to handle your data. Now you'll pull your data into Cogynt to take advantage of the platform's capabilities. This process has two phases:

  1. Deploying your model.
  2. Ingesting your data into Workstation.

Deploy Your Model

Once you've developed a model, you can put it into execution to view the outcome of its patterns, or check whether that the model is working the way you expected. This process is called "deploying" a model. For more information, see Initiating Deployments in the Cogynt Authoring User Guide.

You can deploy your model from several different places within Cogynt Authoring:

Click any of the preceding links to see the corresponding instructions in the Cogynt Authoring User Guide. Although each process begins from a different place, they all involve the same fundamental steps:

  1. Click Deploy to initiate the deployment.
  2. In the Confirm Deployment dialog:
    1. Check the Reset published topics? checkbox to reset the information gathered from published topics, and give the deployment a clean data set. TBD: WE WANT THIS UNCHECKED FOR THIS EXERCISE, RIGHT?
    2. Click Deploy to confirm the deployment.

Next, you'll use the deployed model to ingest data into Cogynt Workstation so that you can work with your model's results.

Ingest Data into Workstation

Before analysts can begin working with data, the data must be ingested into Cogynt Workstation.

Data ingestion creates a list of unique and total records, applies any event decorations used to make specific events from a stream stand out, and provides the status for that data stream.

Example

Here is a quick example to clarify unique and total records:

Our onboarding dataset contains both comments, and records of accounts making those comments. Each time we update our data set, the total number of records increases (both comments and users making them), but the unique user records remain the same. The unique records therefore can never exceed the total number of records.

Cogynt Workstation's Admin tab, accessible from the Workstation home page, houses several important options for ingesting data and customizing the look and feel of Cogynt.

To ingest data:

  1. On the left side of the Admin screen, locate the Ingestion Settings column. This column lists each event created in Cogynt Authoring sorted by Type or Ingestion status.
  2. In the right-most column in Ingestion Settings, toggle the Status switch to ingest that data.

Ingestion Status

An event's ingestion status indicates the state of a data stream. The possible statuses are as follows:

StatusColorDescription
InactiveGrayThis event type has not been ingested.
RunningGreenThis event type is actively being ingested into Workstation from Kafka. Analysts are able to see events in the event stream.
SuspendingYellow with hash marksThe live ingestion of this event type is in the process of being suspended. Any events already ingested will remain in the system unless deleted.
SuspendedYellowThe live ingestion of this event type has been suspended. Any events already ingested will remain in the system unless deleted.
Topic Does Not ExistRedThe Kafka topic that this event type is consumed from can no longer be found. (This usually indicates the Kafka topic was deleted.)

Next Steps

Once data is ingested, you can move on to viewing it in Workstation.

Analysts and Admins may have different workflows while using Workstation. Here, we've sketched what a typical workflow for each role might look like.

Analysts create and study views:

  1. Add widgets to each view.
  2. Adjust widget placement and size within a view to create a workable space.
    • More than one view may be helpful for complex analysis.
    • At the top of any view, click the tab of a view you wish to switch to.
  3. Drag widgets into other widgets for further analysis.

Admins create Custom Field Templates, Tags, and Event Detail Templates to further standardize or customize workflows:

  1. Custom Field Templates bring additional fields into an event, such as a checklist of items, a dropdown menu, or rich or plain text fields. Analysts can add detailed analysis on an event, compose rich reports of 50,000 characters, and add analyst-specific feedback fields.
  2. Tags color code and help improve the searchability of an event notification or collection.
  3. Event Detail Templates allow administrators to adjust the default layout of an event. Admins can control how information is conveyed to analysts to help emphasize the most important details associated with an event, or customize the spacing or layout of elements such as maps.

For guidelines on putting your own data to work using these processes, see Viewing Your Data.

Did you find the information that you needed?