> For the complete documentation index, see [llms.txt](https://multiset.gitbook.io/multiset/llms.txt). Markdown versions of documentation pages are available by appending `.md` to page URLs; this page is available as [Markdown](https://multiset.gitbook.io/multiset/fundamentals/images-and-media.md).
# Analytics and Usage
The **Analytics** section of the MultiSet Developer Portal reports how your maps and objects perform in production: how many localization requests they receive, how many of those return a pose, and where inside a map the requests are coming from.
Analytics is split into three tabs, **Overview**, **Maps Query**, and **Object Query**. Each one can be scoped with the date filters at the top right (Last 7 Days, Last 30 Days, Last 90 Days, or a custom range).
## Overview
The Overview tab aggregates every localization request across your account for the selected period.
Overview: account-wide query summary with a per-map breakdown
The Query Summary reports four values:
| Metric | Meaning |
| -------------- | ----------------------------------------------------------- |
| Total Queries | Every localization request received in the selected period. |
| Pose Found | Requests that returned a pose. |
| Pose Not Found | Requests where no pose could be estimated. |
| Recall Rate | Pose Found divided by Total Queries. |
Below the tiles, two donut charts split the same totals into **Maps** and **Object Tracking**, each with its own found count and recall percentage, so you can see at a glance whether map localization or object tracking is underperforming.
The **Queries by Map** table then breaks the totals down map by map. For every map it lists the Map Name, Map Code, query count, Pose Found (with that map's own recall percentage), Pose Not Found, and the timestamp of its most recent query.
## Maps Query
The Maps Query tab scopes everything to one map. Pick a map from the dropdown to get its own Query Summary, recall rate, and donut chart, which makes it straightforward to compare how reliably different spaces localize.
Maps Query: recall rate for a single map
Every localization request against the map is recorded individually with its **confidence value** and its **timestamp**. That per-request detail is what makes the aggregate recall rate actionable: a map sitting at 65% recall can mean poor coverage across the whole space, or it can mean a handful of failing sections dragging down an otherwise reliable map, and the request-level confidence and timing tell you which.
{% hint style="info" %}
Recall rate is a measure of your **map**, not of the SDK. A low rate usually points to sections of the space that were scanned thinly, have changed since the scan, or are visually repetitive. Localization hints such as [HintPosition](/multiset/fundamentals/localization/pose-prior-hintposition.md) and [HintFloorHeight](/multiset/fundamentals/localization/hint-floor-height.md) can also lift recall in large or multi-floor maps.
{% endhint %}
## Localization heatmap
Each map also has a spatial view of its request volume. Select **View Heatmap** on the Maps Query tab to render request density directly onto the map's 3D mesh.
Localization heatmap: request intensity across sections of the map
Warmer colors mark the sections receiving the most localization requests, cooler colors the least. Fly through the mesh with the on-screen controls (WASD to move, E and Q for up and down, mouse drag to orbit) to inspect specific areas.
The heatmap answers questions the aggregate numbers cannot:
* **Where are users actually localizing?** Concentrated hotspots usually mark entrances, lobbies, and decision points, which are the areas worth scanning most thoroughly.
* **Which sections need a rescan?** A dense hotspot in a map with low recall points straight at the part of the space that is failing.
* **Is any of the map unused?** Areas with no requests at all may not need to be maintained in future scans.
## Object Query
The Object Query tab reports the same analytics for object tracking. Scoped to an object, it gives you Total Queries, Pose Found, Pose Not Found, and Recall Rate for that object, with each request recorded with its own confidence value and timestamp, exactly as map queries are.
Because object tracking resolves a pose relative to a single registered object rather than a scanned space, object analytics are reported per object code and have no spatial heatmap.
## Plan usage
Storage and API call allowances for your account are shown under **Account → Plan**.
---
# Agent Instructions
This documentation is published with GitBook. GitBook is the documentation platform designed so that both humans and AI agents can read, navigate, and reason over technical content effectively. Learn more at gitbook.com.
## Querying This Documentation
If you need additional information that is not directly available in this page, you can query the documentation dynamically by asking a question.
Perform an HTTP GET request on the current page URL with the `ask` query parameter, and the optional `goal` query parameter:
```
GET https://multiset.gitbook.io/multiset/fundamentals/images-and-media.md?ask=&goal=
```
`ask` is the immediate question: it should be specific, self-contained, and written in natural language.
`goal` is optional and describes the broader end goal you are ultimately trying to accomplish on behalf of the user. GitBook uses it to tailor the answer towards what is most useful for that goal.
The response will contain a direct answer to the question and relevant excerpts and sources from the documentation.
Use this mechanism when the answer is not explicitly present in the current page, you need clarification or additional context, or you want to retrieve related documentation sections.