> For the complete documentation index, see [llms.txt](https://docs.everesteer.ai/llms.txt). Markdown versions of documentation pages are available by appending `.md` to page URLs; this page is available as [Markdown](https://docs.everesteer.ai/data/column-definitions.md).

# Column definitions

Every structural column, the feature pattern, and the target columns.

The dataset is a single parquet file per split with the same column layout. The live column list is served by `eiq_features.json` and by `get_dataset_schema` on the API and MCP surfaces.

## Structural columns

| Column      | Type   | Description                                                                                                                                                                                                                                                                                                                                                                                |
| ----------- | ------ | ------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------ |
| `id`        | string | Row identifier, unique per row per exped. In the tournament dataset it is 16 hex characters, for example `93a7b0f18e26f106`; another dataset can use a different shape, so copy what you were served and never construct or pattern-match one. Join your predictions to rows on this, within one exped. It never matches an id from another exped or another round. Do not hardcode an id. |
| `exped`     | string | As-of-day identifier, for example `exped_0850`. One cross-section: every row that was knowable on that day, so one exped can span more than one trading date. Sample by this.                                                                                                                                                                                                              |
| `data_type` | string | Which split the row belongs to: `train`, `validation` or `live`.                                                                                                                                                                                                                                                                                                                           |

In the served parquet file `id` is the index, not a column. When you load it with `pd.read_parquet`, check whether it arrives as a column or as the index. The example scripts handle both cases.

## Features

| Pattern          | Type   | Description                                                                         |
| ---------------- | ------ | ----------------------------------------------------------------------------------- |
| `feature_<Name>` | `int8` | Decile bins `0` to `9`. `-1` means the source was unavailable: treat it as missing. |

Feature names are consistent across splits. The live feature list is in `eiq_features.json`. See [Feature groups](/data/feature-groups.md).

## Targets

| Column           | Type      | Description                                                                                                                                                      |
| ---------------- | --------- | ---------------------------------------------------------------------------------------------------------------------------------------------------------------- |
| `target_everest` | `float32` | Primary payout target. 20-day forward return, rank-based, binned into five levels. Values: `0.0`, `0.25`, `0.5`, `0.75`, `1.0`.                                  |
| `target_<Name>`  | `float32` | Auxiliary targets with codenamed suffixes: related forward-looking constructions at several horizons. The name carries neither the horizon nor the construction. |

The table above is the tournament dataset's layout. Its primary target is `target_everest`, and Atlas publishes no `target` alias. The full target list is in `eiq_features.json`. See [Targets](/data/targets.md) for how to use them.

Another dataset, such as the one an event serves, publishes its own target block under its own names and may carry no `target` alias at all. The column you are graded on is always `primary_target` on the `get_dataset_schema` response for the tree you were served. Read it there rather than carrying a name across from this page.

In the `train` split, every target column carries values. In `validation` and `live`, the target columns are present but every value is `NaN` (blanked). A target value of `NaN` means the forward return was uncomputable for that row: drop those rows when training on that target.


---

# 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.everesteer.ai/data/column-definitions.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.