> For the complete documentation index, see [llms.txt](https://docs.powermonitor.com.br/llms.txt). Markdown versions of documentation pages are available by appending `.md` to page URLs; this page is available as [Markdown](https://docs.powermonitor.com.br/en/power-monitor/governanca/dados-e-engenharia/ciencia-de-dados.md).

# Data Science

Inventory of Microsoft Fabric data science and AI artifacts, ML Models, Data Agents, ML Experiments, Ontologies and Graph Models.

The **Data Science** screen catalogs the artificial intelligence, experimentation and predictive modeling artifacts published in your Microsoft Fabric tenant, with the workspace of each one and who created and modified them.

**How to access:** menu *Governance › Data and Engineering › Data Science*. The screen is read-only and is available to all user profiles (respecting the workspace scope and page blocking configured for your user).

<figure><picture><source srcset="/files/UTKDFLkBOCC0o0qiwzNC" media="(prefers-color-scheme: dark)"><img src="https://3938213054-files.gitbook.io/~/files/v0/b/gitbook-x-prod.appspot.com/o/spaces%2FH2bFRBmIfyK3kwVKbldl%2Fuploads%2Fgit-blob-4ee502feead752061bf99381ae9c002e4ecc8bfd%2Fpm-governanca-ciencia-de-dados-en.png?alt=media" alt="Data Science screen with the cards by type and the unified table"></picture><figcaption><p>Governance › Data and Engineering › Data Science</p></figcaption></figure>

## What it is for

* **Give visibility to AI assets**: know how many ML models, data agents (Data Agents) and experiments exist and where they are.
* **Avoid scattered experiments**: identify experiments and models spread across development or test workspaces.
* **Audit owners**: find out who created or modified a Data Agent or a model before promoting or retiring it.
* **Review access**: check, through the details modal, who has access to each artifact.

## Screen overview

The screen has three blocks, from top to bottom:

1. **Cards by artifact type**: counters that also work as a type filter.
2. **Period filters** (**Created at** and **Modified at**), just above the table.
3. **Inventory table**, with text filters under the column titles, sorting, pagination and the **⋮** actions menu on each row.

While the data loads, the message **Loading...** appears above the table; the five types are loaded in parallel, so the table may fill in gradually. When no item matches the filters, the table shows **No data science artifacts found**.

## Features

### Cards by artifact type

**What it is:** a row of six cards (**All**, **ML Model**, **Data Agent**, **ML Experiment**, **Ontology** and **Graph Model**) each with the number of items of that type.

**What it is for:** seeing the size of the AI inventory at a glance and isolating a type with one click (for example, reviewing only the Data Agents).

<figure><picture><source srcset="/files/jCCBQaLkLLUaiQmwJbfs" media="(prefers-color-scheme: dark)"><img src="https://3938213054-files.gitbook.io/~/files/v0/b/gitbook-x-prod.appspot.com/o/spaces%2FH2bFRBmIfyK3kwVKbldl%2Fuploads%2Fgit-blob-1379cb500188be96adcc49733da871bdaaaeeaf2%2Fpm-governanca-ciencia-de-dados-cards-en.png?alt=media" alt="All, ML Model, Data Agent, ML Experiment, Ontology and Graph Model cards"></picture><figcaption><p>Cards by artifact type</p></figcaption></figure>

**How to use:**

1. In *Governance › Data and Engineering › Data Science*, click the card of the desired type (for example, **Data Agent**).
2. The card is highlighted and the table shows only the items of that type, returning to the first page.
3. To return to the complete list, click the **All** card.

<figure><picture><source srcset="/files/w8rwTRkRLIitIcWeqjgW" media="(prefers-color-scheme: dark)"><img src="https://3938213054-files.gitbook.io/~/files/v0/b/gitbook-x-prod.appspot.com/o/spaces%2FH2bFRBmIfyK3kwVKbldl%2Fuploads%2Fgit-blob-56ee995e5dd83be049ee8840603a16c512c8ed32%2Fpm-governanca-ciencia-de-dados-filtro-tipo-en.png?alt=media" alt="Data Agent card selected and table filtered to that type"></picture><figcaption><p>Data Agent card selected, filtering the table</p></figcaption></figure>

**How it works:**

| Card              | What it counts                                      |
| ----------------- | --------------------------------------------------- |
| **All**           | Total of the five types below                       |
| **ML Model**      | Machine Learning models registered as a Fabric item |
| **Data Agent**    | Data agents (Fabric Data Agent)                     |
| **ML Experiment** | Machine Learning experiments                        |
| **Ontology**      | Ontologies                                          |
| **Graph Model**   | Graph models                                        |

* The counters show the total of each type in the monitored workspaces, **without considering the column filters**: filtering the table does not change the numbers on the cards.
* Text and period filters already filled in remain in effect when you switch cards.

{% hint style="info" %}
To see the **structure** of each ontology (entities, relationships, data bindings, rules and metrics) and the problems found, such as the ontologies still on the old experience, use the [Ontologies](/en/power-monitor/governanca/dados-e-engenharia/ontologias.md) screen.
{% endhint %}

### Text filters by column

**What it is:** text boxes just below the titles of the **Name** (*Item name*), **Workspace** (*Workspace name or ID*), **Created by** and **Modified by** columns.

**What it is for:** quickly locating an artifact by name, by workspace (for example, all items in a production workspace) or by author.

<figure><picture><source srcset="/files/wRchaJF35BGKXgNrb5rD" media="(prefers-color-scheme: dark)"><img src="https://3938213054-files.gitbook.io/~/files/v0/b/gitbook-x-prod.appspot.com/o/spaces%2FH2bFRBmIfyK3kwVKbldl%2Fuploads%2Fgit-blob-a5a1bcab58c91fb5b312a621c96a8f660d17daf8%2Fpm-governanca-ciencia-de-dados-filtros-coluna-en.png?alt=media" alt="Table header with the filter boxes by column"></picture><figcaption><p>Text filters under the column titles</p></figcaption></figure>

**How to use:**

1. Type the term in the filter box of the desired column.
2. The table is filtered automatically, a few moments after you stop typing, and returns to the first page.
3. To remove the filter, delete the text in the box.

**How it works:**

* The search matches **part** of the text and is **not case-sensitive**.
* In the **Workspace** column, you can type the workspace name or ID.
* Filters on different columns are combined (an item must match all of them).

### Period filters (Created at and Modified at)

**What it is:** two date range selectors, **Created at** and **Modified at**, displayed above the table, each with a start date and an end date (*… to …*).

**What it is for:** answering questions such as "which Data Agents were created this quarter?" or "what has changed since the last review?".

<figure><picture><source srcset="/files/4TXDWrQlZIcQA2n2PJZZ" media="(prefers-color-scheme: dark)"><img src="https://3938213054-files.gitbook.io/~/files/v0/b/gitbook-x-prod.appspot.com/o/spaces%2FH2bFRBmIfyK3kwVKbldl%2Fuploads%2Fgit-blob-3e97972460e0328c31183db8e01a910043bbe87f%2Fpm-governanca-armazenamento-filtros-data-en.png?alt=media" alt="Created at and Modified at period selectors above the table"></picture><figcaption><p>Period filters</p></figcaption></figure>

**How to use:**

1. In the **Created at** (or **Modified at**) selector, fill in the start date, the end date or both.
2. The table shows only the items within the range.
3. To remove the filter, clear the dates.

**How it works:**

* The range includes the entire start day (from 00:00) and the entire end day (until 23:59), in your browser's time zone.
* Items without the corresponding date are removed from the list while the filter is filled in.
* When you fill in a range, the table is re-sorted by the corresponding date: from oldest to newest when there is a start date; from newest to oldest when there is only an end date.

### Sorting

**What it is:** sorting the table by clicking a column title.

**What it is for:** seeing the most recently modified items first, or grouping by workspace or by author.

**How to use:**

1. Click the column title (for example, **Modified at**). The first click sorts in **descending** order.
2. Click the same title again to reverse the order. An icon next to the title indicates the sorted column and direction.

**How it works:** the default sort is by **Modified at**, from newest to oldest. The **Type** column and the actions column are not sortable.

### Inventory table

**What it is:** a single table that lists the five artifact types.

**What it is for:** checking, in one place, where each AI artifact is and who last worked on it.

**How it works:**

| Column          | Description                                                                                                                         |
| --------------- | ----------------------------------------------------------------------------------------------------------------------------------- |
| **Type**        | Colored badge with the icon and name of the type (**ML Model**, **Data Agent**, **ML Experiment**, **Ontology** or **Graph Model**) |
| **Name**        | Artifact name (**No name** when Fabric does not provide it)                                                                         |
| **Workspace**   | Workspace where the item is published                                                                                               |
| **Created by**  | Who created the item                                                                                                                |
| **Modified by** | Who made the last change                                                                                                            |
| **Created at**  | Creation date                                                                                                                       |
| **Modified at** | Date of the last modification                                                                                                       |

When Fabric does not provide a value, the cell displays "-" (or **Not reported**, in the date columns).

### Pagination

**What it is:** navigation between pages in the table footer, which displays 10 items per page.

**What it is for:** browsing large inventories without overloading the screen.

<figure><picture><source srcset="/files/zbYKEIuqwDHknagIDxix" media="(prefers-color-scheme: dark)"><img src="https://3938213054-files.gitbook.io/~/files/v0/b/gitbook-x-prod.appspot.com/o/spaces%2FH2bFRBmIfyK3kwVKbldl%2Fuploads%2Fgit-blob-89b549109009a731f242253ddf454ee42bf2c99a%2Fpm-governanca-ciencia-de-dados-paginacao-en.png?alt=media" alt="Table footer with page navigation"></picture><figcaption><p>Table pagination</p></figcaption></figure>

**How to use:** use the footer controls to move forward, back or go to a specific page. Switching cards or changing a filter takes you back to the first page.

### Row actions menu (⋮)

**What it is:** the **⋮** button at the end of each row (or right-clicking the row) opens the menu with the **Open details** item, which has a submenu with the modal tabs.

**What it is for:** going straight to the tab you need, without going through the Overview.

<figure><picture><source srcset="/files/4VR7q3bOXCOI93TpUrt5" media="(prefers-color-scheme: dark)"><img src="https://3938213054-files.gitbook.io/~/files/v0/b/gitbook-x-prod.appspot.com/o/spaces%2FH2bFRBmIfyK3kwVKbldl%2Fuploads%2Fgit-blob-4733e8b10d3dd90950aa1f399ebd5319f4c40e66%2Fpm-governanca-ciencia-de-dados-menu-acoes-en.png?alt=media" alt="Row actions menu with the Open details submenu"></picture><figcaption><p>Actions menu with the submenu of modal tabs</p></figcaption></figure>

**How to use:**

1. Click the row's **⋮** (or right-click it).
2. Click **Open details** to open the modal on the **Overview** tab, **or** hover over **Open details** and choose **Overview** or **Users & Permissions** in the submenu.

Clicking anywhere on the row also opens the details modal on the **Overview** tab.

### Details modal: Overview tab

**What it is:** the artifact's record panel. The header shows the name, a badge with the type and the workspace.

**What it is for:** checking the item's data and copying the identifier to use in the Fabric portal, in scripts or in support tickets.

<figure><picture><source srcset="/files/mST2epXRBxioIhNXV9ZY" media="(prefers-color-scheme: dark)"><img src="https://3938213054-files.gitbook.io/~/files/v0/b/gitbook-x-prod.appspot.com/o/spaces%2FH2bFRBmIfyK3kwVKbldl%2Fuploads%2Fgit-blob-ff843747ac7433bebdc3d2ded4eb2aa0deaa4ca1%2Fpm-governanca-ciencia-de-dados-modal-visao-geral-en.png?alt=media" alt="Details modal of a data science artifact on the Overview tab"></picture><figcaption><p>Details modal: Overview tab</p></figcaption></figure>

**How to use:**

1. Click the artifact row (or **⋮** › **Open details**).
2. Check the **ID**, **Name**, **Type**, **Workspace**, **Created at**, **Modified at**, **Created by**, **Modified by** fields and, when present, **Tags**.
3. To copy the identifier, click **Copy** next to the **ID**; the button changes to **Copied!**.
4. To close, click the **X** in the header or outside the modal.

**How it works:** fields without a value in Fabric are not displayed. The modal for these types has only the **Overview** and **Users & Permissions** tabs: there are no dependencies or schedules tabs, because Fabric does not declare dependencies for them.

### Details modal: Users & Permissions tab

**What it is:** the list of users, groups and applications with access to the artifact, with the total at the top (*N users and applications with access*) and the counter also in the tab title.

**What it is for:** reviewing who can use or change a Data Agent or a model, including external guests, before promoting it to production.

<figure><picture><source srcset="/files/D6mbdpMKXqlAM93oTYAV" media="(prefers-color-scheme: dark)"><img src="https://3938213054-files.gitbook.io/~/files/v0/b/gitbook-x-prod.appspot.com/o/spaces%2FH2bFRBmIfyK3kwVKbldl%2Fuploads%2Fgit-blob-232a397c191679d3d0513a0c4df5a27811009366%2Fpm-governanca-ciencia-de-dados-modal-usuarios-en.png?alt=media" alt="Details modal of a Data Agent on the Users and Permissions tab"></picture><figcaption><p>Details modal: Users &#x26; Permissions tab</p></figcaption></figure>

**How to use:**

{% stepper %}
{% step %}

#### Open the tab

Click the row's **⋮** and choose **Open details** › **Users & Permissions** (or open the modal and click the tab).
{% endstep %}

{% step %}

#### Search for a principal

Type in the **Search by name, email or identifier...** box to locate a user, group or application.
{% endstep %}

{% step %}

#### Evaluate the access

Check the columns of each row:

| Column             | Description                            |
| ------------------ | -------------------------------------- |
| **User**           | Name of the user, group or application |
| **Identifier**     | E-mail or identifier of the principal  |
| **Principal Type** | **User**, **Group** or **Application** |
| **Category**       | **Member** or **Guest/External**       |
| **Access Level**   | Permission granted on the item         |
| {% endstep %}      |                                        |
| {% endstepper %}   |                                        |

**How it works:** the list is loaded when the modal opens. If there is no one, **No users found.** appears; if the query fails, **Could not load users. Please try again.** appears (close and reopen the modal).

## Rules and behavior

* **Data source:** artifacts are collected by the [Inventory Scan](/en/power-monitor/mapeamento/inventarios.md). Changes in Fabric appear after the next scan run.
* **Only Fabric items appear.** A model trained inside a notebook is only listed after it is registered as an **ML Model** item in the workspace.
* **Monitored workspaces only:** the list shows the artifacts of the workspaces the organization chose to monitor; there is no option to include unmonitored workspaces.
* **Workspace scope:** if your user has visibility restricted to some workspaces, you will see only their artifacts.
* **Items deleted** in Fabric are removed from this list and appear in [Deleted Artifacts](/en/power-monitor/governanca/operacao/artefatos-excluidos.md).
* **Dates** are displayed in your browser's time zone.
* **Read-only:** the screen does not change anything in Fabric; changes must be made in the portal itself.

## Step by step: audit the published Data Agents

{% stepper %}
{% step %}

#### Filter the type

Click the **Data Agent** card.
{% endstep %}

{% step %}

#### Refine by workspace

Use the **Workspace** column filter to focus, for example, on production workspaces.
{% endstep %}

{% step %}

#### See what changed recently

Fill in the start date of the **Modified at** filter to list the agents changed since the last review.
{% endstep %}

{% step %}

#### Review the access

Open the **⋮** menu of each agent and choose **Open details** › **Users & Permissions** to see who can use it, including external guests (**Guest/External**).
{% endstep %}
{% endstepper %}

## Frequently asked questions

<details>

<summary>A model trained in a notebook does not appear under "ML Model". Why?</summary>

The screen lists Fabric items. The model is only cataloged here after it is formally registered as an item of the **ML Model** type in the workspace.

</details>

<details>

<summary>How often is the list updated?</summary>

On each run of the Inventory Scan. See the scan page to view and adjust the schedule.

</details>

<details>

<summary>I filtered the table, but the number on the cards did not change.</summary>

This is the expected behavior: on this screen the cards show the total of each type, without considering the column filters.

</details>

<details>

<summary>Can I change or delete an artifact on this screen?</summary>

No. The screen is for inventory and lookup; changes must be made in Microsoft Fabric itself.

</details>

## Related pages

* [Data Engineering](/en/power-monitor/governanca/dados-e-engenharia/engenharia-de-dados.md)
* [Ontologies](/en/power-monitor/governanca/dados-e-engenharia/ontologias.md)
* [Development](/en/power-monitor/governanca/dados-e-engenharia/desenvolvimento.md)
* [Real Time](/en/power-monitor/governanca/dados-e-engenharia/tempo-real.md)
* [Deleted Artifacts](/en/power-monitor/governanca/operacao/artefatos-excluidos.md)
* [Mapping › Inventory](/en/power-monitor/mapeamento/inventarios.md)


---

# 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 by asking a question.

Perform an HTTP GET request on the following URL with the `ask` and `goal` query parameters:

```
GET https://docs.powermonitor.com.br/en/power-monitor/governanca/dados-e-engenharia/ciencia-de-dados.md?ask=<question>&goal=<user_goal>
```

`ask` is the immediate question: it should be specific, self-contained, and written in natural language.
`goal` is what the user is ultimately trying to achieve, the reason they need the answer. Sharing it helps GitBook give you a better, more relevant answer. A goal is most helpful when it describes the outcome the user wants rather than restating the question. For example, with `ask=how do I create an API token`, a goal like `automate deployments from our CI pipeline` lets GitBook tailor the answer to that use case.

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.