# Data Overview

Amplitude exposes six entities through Coupler.io, ranging from raw event-level data to aggregated behavioral metrics. The right entity depends on whether you need granular user actions or summary-level analytics.

## Entities at a glance

| Entity                  | Granularity         | Best for                              |
| ----------------------- | ------------------- | ------------------------------------- |
| Events                  | Row per event       | Funnel analysis, user journey mapping |
| Active users            | Row per time period | DAU/WAU/MAU tracking                  |
| Average session lengths | Row per time period | Engagement benchmarking               |
| Cohorts                 | Row per cohort      | Retention and segmentation analysis   |
| Events list             | Row per event type  | Auditing tracked events               |
| Annotations             | Row per annotation  | Contextualizing metric changes        |

## Events

#### Event fields

| Field              | Description                                          |
| ------------------ | ---------------------------------------------------- |
| `event_type`       | Name of the event (e.g., `button_click`, `purchase`) |
| `event_time`       | Timestamp of when the event occurred                 |
| `user_id`          | Identifier for the user who triggered the event      |
| `device_id`        | Device identifier (used when user\_id is absent)     |
| `session_id`       | Groups events into a single session                  |
| `platform`         | Platform (iOS, Android, Web)                         |
| `country`          | Country of the user                                  |
| `region`           | Region/state of the user                             |
| `os_name`          | Operating system                                     |
| `device_type`      | Device category                                      |
| `event_properties` | Key-value pairs specific to the event                |
| `user_properties`  | Key-value pairs describing the user at event time    |

## Active users

#### Active user fields

| Field     | Description                                         |
| --------- | --------------------------------------------------- |
| `date`    | Date of the measurement                             |
| `value`   | Count of active users on that date                  |
| `country` | Country of users (when group-by-country is enabled) |

## Average session lengths

#### Session length fields

| Field   | Description                              |
| ------- | ---------------------------------------- |
| `date`  | Date of the measurement                  |
| `value` | Average session duration in milliseconds |

## Cohorts

#### Cohort fields

| Field           | Description                              |
| --------------- | ---------------------------------------- |
| `id`            | Unique cohort identifier                 |
| `name`          | Name of the cohort                       |
| `size`          | Number of users in the cohort            |
| `description`   | Cohort description                       |
| `last_computed` | Timestamp of the last cohort computation |
| `owners`        | List of owners                           |
| `published`     | Whether the cohort is published          |

## Events list

#### Event type fields

| Field        | Description                                      |
| ------------ | ------------------------------------------------ |
| `event_type` | Name of the tracked event                        |
| `totals`     | Total number of times this event has occurred    |
| `visible`    | Whether the event is visible in the Amplitude UI |

## Annotations

#### Annotation fields

| Field     | Description                    |
| --------- | ------------------------------ |
| `id`      | Annotation identifier          |
| `date`    | Date the annotation applies to |
| `label`   | Short annotation label         |
| `details` | Full annotation text           |

## Common metric combinations

* **Events + Active users** — Join these in Coupler.io to normalize event volume by active user count and calculate per-user engagement rates
* **Events + Cohorts** — Use Cohorts to filter or segment Events data when building retention dashboards
* **Active users over time + Annotations** — Append Annotations alongside Active users data to explain spikes and drops in your charts
* **Average session lengths + Active users** — Combine with an Aggregate transformation to surface engagement quality alongside reach metrics

## Use cases by role

{% tabs %}
{% tab title="Product managers" %}

* Pull **Events** data into BigQuery or Google Sheets to build custom funnel reports outside Amplitude's UI
* Track **DAU/WAU/MAU** trends in a Looker Studio dashboard that refreshes automatically
* Use **Cohorts** to compare behavior between user segments in a spreadsheet pivot table
* Feed event data into ChatGPT or Gemini to ask natural language questions about user behavior patterns
  {% endtab %}

{% tab title="Marketers" %}

* Combine **Active users** (grouped by country) with ad spend data from your ad platform using a Join transformation to calculate cost per active user by region
* Export **Events** to track campaign-triggered actions (signups, purchases) alongside attribution data
* Use **Annotations** to document marketing campaigns directly in your analytics timeline
  {% endtab %}

{% tab title="Data analysts" %}

* Export raw **Events** to BigQuery for custom SQL analysis not available in Amplitude's built-in charts
* Use **Events list** to audit which events are actively tracked and clean up your taxonomy
* Append **Active users** from multiple Amplitude projects into a single dataset using the Append transformation for cross-product comparisons
* Schedule recurring exports to keep a historical data warehouse in sync with Amplitude
  {% endtab %}
  {% endtabs %}

## Platform-specific notes

* The **Events** entity uses Amplitude's Export API, which returns compressed data in batches. Large date ranges on high-volume projects can trigger timeouts — use the **request time range** parameter to break requests into smaller intervals.
* **Active users grouped by country** makes a separate API call per country. If your project tracks many countries, this can be slow or cause connection errors — disable it if needed.
* **Cohort** data reflects the state of the cohort at the time of the sync, not historical membership snapshots.
* The `event_properties` and `user_properties` fields contain nested JSON. Some destinations (like BigQuery) will store these as JSON columns; spreadsheet destinations may flatten or truncate them.
* Amplitude's Export API only supports data from after your project's creation date — you cannot pull data that predates the project.