Guides
Start typing to search…

Event Explore

The Event Explorer (Report View) is a powerful analytics engine designed for deep-dive analysis. It allows you to correlate standard web telemetry with specialized AI agent behaviors, visualize trends, and perform mathematical aggregations on your data.

1. Creating a New Report

Follow these steps to build a custom analytics report from scratch.

Step 1: Report Setup

  • Title: Click on Untitled Report at the top left to rename your report (e.g., "Weekly Sign-ups").
  • Description: Click Add description... to provide context or objectives for this analysis.
  • Controls: Use the buttons in the top right to Save your report or Close the builder.

Step 2: Configuring Data (Left Sidebar)

Use the sidebar to define the logic of your report:

  • Events or Metrics:

  • Click the Select Metric or Event field to choose the primary action to track (e.g., Page View, Button Click, Purchase).

  • You can add multiple metrics (labeled A, B, C...) to compare trends side-by-side.

  • Filter & Breakdown:

  • Click the + icon to open the Property Picker (see "Navigating Property Data" below).

  • Example: Filter by Country is United States or break down by Browser.

  • Apply: Click the Apply button at the bottom of the sidebar to render your data.

2. Navigating Property Data (New)

When you click to add a Filter or Breakdown, the Search Properties modal appears. This interface helps you find the exact data point you need.

Property Categories

The left sidebar organizes hundreds of data points into logical groups:

  • Agent Conversation Property: Specifics about your AI's behavior (e.g., AI agent id, LLM model name).
  • User / Subscription Property: Attributes related to the customer (e.g., Email, Plan Type).
  • Event Property: Context specific to the action (e.g., Button Text, Error Message).
  • Technical Context: URL, Inferred Device, and UTM parameters for tracking sources and environments.

Understanding Data Types

Icons next to each property name indicate how it can be used:

  • Tt (String): Text data (e.g., Agent Name). Best for grouping and breaking down data.
  • # (Number): Numeric data (e.g., Cached tokens, Estimated cost). Best for mathematical aggregations (Sum, Average).

Property Context

Select any property in the list to see a definition panel on the right. This confirms:

  • Description: What the data point actually represents (e.g., "The unique identifier for the specific AI agent configuration").
  • Type: The technical format (e.g., STRING, INTEGER).

3. Advanced: Aggregating Numeric Properties

Beyond simple event counts, you can calculate values like total revenue, latency, or token usage.

  1. Select an Event: Choose an event that carries a numeric value (e.g., Action Called).
  2. Change Aggregation: Click the dropdown that defaults to Unique or Total Events.
  3. Select "Aggregate Property": This switches the metric from a counter to a calculator.
  4. Configure the Math:
  • Property: Select a # (Numeric) property (e.g., Input tokens, Estimated cost).
  • Method: Choose the operation:
  • Sum: Adds all values (e.g., Total Token Usage).
  • Average: Calculates the mean (e.g., Average Latency).
  • Min / Max: Identifies outliers.

4. Visualization Settings (Top Toolbar)

Adjust how the data is displayed on the timeline:

  • Time Range: Select a preset window (7D, 30D, 60D, 90D) or use Custom.
  • Interval: Group data by hour, day, week, or month (currently set to Daily).
  • Chart Type: Switch between visualization styles (e.g., Line chart).

Updated about 2 months ago