Skip to main content

User Activities

The User Activities module provides usage analytics and logs for all model-related operations on the platform. It enables teams to monitor validation and prediction activity, track per-user behavior, and inspect image stream operations.

Sections Overview​

This page is divided into four major sections:

1. Validation Stats​

  • Displays a circular graph representing the total usage of deployed models in validation tasks.
  • Each model type is color-coded.
  • Hovering over each segment reveals exact usage counts.
  • Useful for identifying the most frequently validated models.

2. Single Prediction Stats​

  • Provides a donut chart showing model usage specifically for single prediction operations.
  • Allows you to see which models are popular for on-the-fly analysis.
  • A quick glance helps understand how prediction endpoints are consumed by users.

3. User Activities Log​

This section captures all model-related operations triggered by users, including batch validations and single predictions.

Columns​

  • Deployment ID: Unique deployment or run ID.
  • Model Name: The AI model used in the operation.
  • Dataset: Dataset on which the model was executed.
  • Type: Specifies if it was a Batch Run or Single Prediction.
  • User: Email of the user initiating the action.
  • Time Taken: Duration of execution.
  • Date: Timestamp of the activity.

Pagination controls are available to browse through large activity logs.

4. Image Stream Stats​

This section logs interactions involving streamed image processing.

Columns​

  • Image ID: Unique ID for the streamed image.
  • Dataset: Source dataset for the image.
  • User: Email of the user involved in the stream.
  • Date: When the image activity occurred.

If no image stream activity has taken place, this section will show a placeholder indicating no data.

Why It Matters​

  • Monitor which models are most frequently used and by whom.
  • Detect performance issues based on time taken for execution.
  • Track how datasets are leveraged across users and tasks.
  • Ensure transparency and traceability across predictive workflows.