Skip to main content

Using the Data Analyzer AI

Beta This is a beta feature, which means we’re still developing it. Some functionality might change.

Before you begin using the Data Analyzer AI, you must install the Data Analyzer Template App and the Data Analyzer plugin. See Data Analyzer Installation to get started.

The Data Analyzer Dashboard

On your Data Analyzer Dashboard, you'll see:

DA.dashboard.png

From the dashboard you can filter by adjusting Model Created Date, Table Name, and Outcome and drill down on the results. If you're using the Data Analyzer for the first time, we strongly recommend that you explore the sample data and models. 

Steps to generate a Model (Insights)

  1. Go to the Data Analyzer app with an installed plugin
  2. From the main dashboard or from the Models table click New Model button.
  3. In the form, click Start the wizard
  4. Fill in all necessary data in the wizard
  5. Wait for the model to be created or you can close the window without stoping the process.
  6. Open the created model and read the insights

Steps to deploy a Model

If you are happy with the model generated, you may decide to “deploy” it and it will dynamically calculate the prediction score for every record in your data.

  1. Open the model
  2. Go to Model Deployment
  3. Copy the Prediction Formula
  4. Go to original table containing the source data that was analyzed
  5. Create a new percentage formula field
  6. Paste the formula

Steps to change the Prediction Type

You may want a simple YES/NO prediction instead of percentage in your source table prediction score.

  1. Edit the model record
  2. Go to Model Deployment section
  3. Change the prediction type from Percentage to YES/NO and Save the changes. Don’t worry, you can change it back.
  4. The Prediction Formula will now give a YES/NO prediction rather than percentage. The formula field in the source table needs to be a “text” one.

Optional: Change the Probability Threshold to the desired percentage and click Save. Default value is 50%, meaning that if the prediction formula gives a prediction value for the record of 49%, it would result in “NO” and 51% would result in YES.

Requirements

Requirements for the data you load in the Data Analyzer:

  • Data size
    • at least 80 records
    • At least 5 fields (without the prediction field)
    • up to 1GB
    • Up to 2M records
  • Outcome requirements
    • Exact value - checked/unchecked and text.
    • Case-sensitive.
    • Needs to occur in 2% to 80% of the records
    • You can build models on up to 10 unique outcome fields for the whole realm.

Related articles

Top