> For the complete documentation index, see [llms.txt](https://docs.neurolabai.io/llms.txt). Markdown versions of documentation pages are available by appending `.md` to page URLs; this page is available as [Markdown](https://docs.neurolabai.io/start-using-neurolab/section-inspector-and-data-explorer.md).

# Section Inspector and Data Explorer

These two available beta views help answer focused questions without changing the report. Section Inspector is the visual, evidence-aware reading surface. Data Explorer turns natural-language requests into bounded operations over the same immutable result.

## Section Inspector

Open **Section Inspector** and choose **Mechanism of Action**, **Cognition**, or **Side Effects**. Each supported section can show:

* the report’s original ranking and raw score;
* a within-section magnitude bar for visual comparison;
* a plain-language interpretation;
* reliability, display-mode, and scope warnings;
* evidence returned with the report;
* separately grouped retrieved literature with verified citation metadata; and
* shortcuts to Raw Report, Knowledge Graph, or report-grounded chat.

The bar is normalized for display within the selected list. It is not a calibrated probability, effect size, or cross-section scale. The rank and raw score remain visible so the visualization does not replace the data.

### Section-specific boundaries

* **MoA:** a ranked class is a model output, while receptor facts and literature are separate evidence states.
* **Cognition:** the visualization can be withheld when the report marks the head out of scope or unsuitable for top-1 emphasis.
* **Side effects:** the shortlist is not an estimate of individual risk, incidence, severity, contraindication, or completeness.

External literature is presented after report and KG material and remains visibly separate. A retrieved paper does not retroactively validate every ranked item.

## Data Explorer

Data Explorer accepts a natural-language instruction, but it does not send that text to the prediction API as an open-ended query. The application chooses one bounded action and sends a structured, read-only request.

Supported intents include:

| Intent                | Effect                                                        |
| --------------------- | ------------------------------------------------------------- |
| Show top candidates   | Return a bounded prefix of the existing ranking.              |
| Filter candidates     | Match text against existing labels, case-insensitively.       |
| Compare candidates    | Compare two to four candidates already present in the report. |
| Highlight a candidate | Focus an existing candidate without changing its rank.        |
| Inspect evidence      | Retrieve candidate-specific evidence when available.          |
| Reset                 | Return the view to its initial state.                         |

Examples of useful requests include:

* “Show the top five mechanisms.”
* “Filter side effects for labels containing sleep.”
* “Compare the first three cognition candidates.”
* “Highlight the second-ranked mechanism.”
* “Inspect the evidence attached to this candidate.”

The natural-language layer chooses an allowed operation and makes its effect inspectable. If a request cannot be represented safely within the available actions, it should fail or ask for a narrower instruction rather than execute arbitrary analysis.

The service accepts only `moa_top_k`, `cog_top_k`, or `se_top_k`; only `ranked_candidates` or `candidate_evidence`; and a maximum result limit of 50. It does not accept SQL, arbitrary field paths, mutations, or free-form execution.

## What remains invariant

Data Explorer does not retrain the model, rerank the report, add a candidate, change a score, fill a withheld section, or turn contextual evidence into field-specific support. Think of the Data Explorer as a controlled lens over a fixed artifact.

The developing [Data Studio](/organize-and-compare-research/data-studio.md) builds on this interaction model for broader analysis while retaining the same provenance and immutability principles.

***

Next: [Knowledge Graph Explorer](/start-using-neurolab/knowledge-graph-explorer.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 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://docs.neurolabai.io/start-using-neurolab/section-inspector-and-data-explorer.md?ask=<question>&goal=<endgoal>
```

`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.