> For the complete documentation index, see [llms.txt](https://arcstar.gitbook.io/arcstar-dao/llms.txt). Markdown versions of documentation pages are available by appending `.md` to page URLs; this page is available as [Markdown](https://arcstar.gitbook.io/arcstar-dao/tournament/data.md).

# Data

## Dataset

<figure><img src="/files/6IF9YU51K0gc3yVeZxni" alt=""><figcaption></figcaption></figure>

Arcstar provides its data scientist community free curated, high quality and obfuscated data.

There are 6 datasets in the Tournament.

| X\_train | ​/data/X\_train.csv​ |
| -------- | -------------------- |
| y\_train | ​/data/y\_train.csv​ |
| X\_test  | ​/data/X\_test.csv​  |

### ID <a href="#id" id="id"></a>

Each `id` in `X_train and X_test` corresponds to a stock at a specific time `Moons`.

### Moons <a href="#moons" id="moons"></a>

The frequency of the `Moons depends on the dataset :`

* gordon-geeko : 30 days interval between each moon
* dolly : 90 days interval
* e-kinetic : 7 days interval
* c-mechanics : 7 days interval
* b-volatility : 7 days interval
* 3b1-signal : 7 days interval

### Features <a href="#features" id="features"></a>

The `features` describe specific attributes of a stock at a point in time.

### Targets <a href="#targets" id="targets"></a>

The `y_train` file contains 3 targets `target_r`, `target_g`, `target_b` that correspond to the idiosyncratic return of the stock over 3 time horizons : 30, 60 and 90 days respectively.

### Split <a href="#split" id="split"></a>

The overall dataset is splitted in two : train and test. The test set start one moon after the last moon of X\_train.

Files might be big (200+MB) so make sure to have enough space before downloading.

### ​ <a href="#undefined" id="undefined"></a>