### Main Contest

This competition is evaluated on the **macro F1 score**. 

F1 Score is calculated as:

$$ F_{1} = 2 \times \frac{precision \times recall}{precision + recall}$$

In the event of a tie, the first to submit their response will be awarded the prize.

### Etiq Contest

The most points win given the following scoring criteria:

* 1 point for running lineage
* 1 point for a clean lineage (No hanging nodes)
* 1 point for running tests (include results in comments or markdown cells for validation)
* 1 point for using RCA (include results in comments or markdown cells for validation)
* 1 point per new bug found per submission (given what Etiq supports, see below)

In the event of a tie, the first to submit their response will be awarded the prize.


**Etiq Supports**:

* Python 3.10-3.13
* Mac OS - ARM
* Windows - Intel/AMD64
* Linux - Intel/AMD64
* Libraries:
 * Pandas
 * Polars
 * Scikit-Learn
 * Numpy
 * Spark


## Submission File

### Main Contest
Submissions must be in a CSV file with two columns: ```AccidentId``` and ```Gravity```. The Gravity column should contain your predictions, either "Lethal" or "NonLethal".

    AccidentId,Gravity
    517754,"Lethal"
    517755,"NonLethal"
    517756,"NonLethal"
    etc.

### Etiq Contest

Start a discussion with the title Etiq contest submission, please include a list of the bugs you have found and upload the notebook with your lineage, tests and RCA, for the bug consideration upload the script/notebook and data you used.

You will not be able to use Etiq directly in the Kaggle Notebook, but if you are interested in this side competition you can use it as a jupyter or vscode extension.