Skip to main content
Version: 2.14.X

Cogynt for Data Scientists

As a data scientist, you will most likely use Cogynt to design models for processing and interpreting large amounts of data. Afterward, you will likely pass off the model's findings to a team of analysts and case managers for review and analysis.

In this capacity, your main tasks in Cogynt include:

  1. Planning your project.
  2. Uploading the relevant data into Cogynt.
  3. Developing and deploying a model with Cogynt Authoring.

Each of these tasks is outlined in greater detail in the sections to follow. Where applicable, links to the appropriate pages in Cogynt's documentation are provided.

Project Planning​

The first steps to using Cogynt effectively are to think through the purpose you want Cogynt to serve and to understand the data that your Cogynt instance will process.

  1. Identify your objective: Articulate the problem you want to solve with Cogynt or the objective that you want Cogynt to fulfill.
  2. Access domain expertise: Ensure that you have some domain knowledge in the area of your problem/objective; otherwise, work closely with a subject matter expert who understands the problem/objective.
  3. Obtain data and data sources: Identify and acquire all the data you need to solve your issue.
  4. Clean your data: Ensure that your data has been cleansed of errors, inconsistencies, corruptions, duplications, or missing entries.
  5. Understand your data: Have a sense of how all your data relate to each other. This will help you construct an effective model in Authoring.

Uploading Data into Cogynt​

When your problem is identified and your data is ready, the next step is to give Cogynt access to your data.

  1. Prepare your data: Ensure that the data is properly formatted with the correct datatypes. This will allow Cogynt to read the data accurately. For more information, see Transforming Data in Modeling Best Practices.
  2. Create a project: Make a project to contain your model. For more information, see Creating and Opening Projects in the Cogynt Authoring User Guide.
  3. Get your data into Kafka: Configure a deployment target in Cogynt Authoring, then establish data source connections or upload any necessary data files using the Data Management Tool.

Modeling and Deployment in Authoring​

Once your data is in Cogynt, the next step is to build and deploy a model using Cogynt Authoring.

  1. Create event types that reference your data: Make event types to inform Cogynt what your data represents and how it should be considered. For more information, see Managing Event Types in the Cogynt Authoring User Guide.
  2. Create patterns: Define the logic that your model will use to process and evaluate data. For more information, see Developing Models in the Cogynt Authoring User Guide.
  3. Create a deployment for your model: Define the parameters under which your model will run. For more information, see Configuring Deployments in the Cogynt Authoring User Guide.
  4. Deploy your model: Put your model into execution. For more information, see Initiating Deployments in the Cogynt Authoring User Guide.

Wrapping Up​

Once your model is deployed, your team's analysts and case managers can use Cogynt Workstation to study your model's findings.

Based on your team's feedback, you may wish to revise and refine your model. For advice on continued model development, see Modeling Best Practices.