kaggle

Download competition files

$ kaggle competitions download -c [competition-name]

$ kaggle competitions submit -c [competition-name] -f [submission.csv] -m "[message]"

$ kaggle datasets list -s "[search-term]"

$ kaggle datasets download -d [owner/dataset-name]

$ kaggle competitions leaderboard [competition-name]

$ kaggle models download -m [owner/model-name/framework/variation]

$ kaggle datasets create -p [path/to/dataset]

$ kaggle config view

kaggle command subcommand [options]

kaggle is the CLI for Kaggle, a data science competition platform. It manages competition data, datasets, notebooks, and models.
Authentication requires API credentials from kaggle.com/account. Place kaggle.json in ~/.kaggle/ with proper permissions (chmod 600).
Competition workflow: download gets data files, work locally, submit uploads predictions. leaderboard shows standings; submissions lists your entries.
Datasets are versioned. download gets current version; create initializes new dataset; version adds updates. init generates dataset-metadata.json for new datasets.
Kernels (notebooks) can be pushed to Kaggle for cloud execution. pull downloads notebooks; push uploads with kernel-metadata.json defining runtime settings.

-c, --competition name

Competition name.

Dataset in owner/dataset format.

Submission message.

File to submit or upload.

Local directory path.

Download output directory.

Unzip downloaded files.

Search term for listings.

Sort results by field.

competitions list|files|download|submit|submissions|leaderboard

Manage competition data and submissions.

Browse and manage datasets.

Manage notebooks (kernels).

Access Kaggle models.

Manage configuration.

Requires Kaggle account and API key. Some competitions have acceptance terms. Large datasets may take time to download. API rate limits apply. Dataset uploads need metadata file.

Kaggle was founded by Anthony Goldbloom and Ben Hamner in 2010 as a platform for predictive modeling competitions. Acquired by Google in 2017. The CLI was added to enable programmatic access to platform resources. Kaggle expanded to include datasets, notebooks, and courses. It hosts one of the largest data science communities with millions of members.

dvc(1), mlflow(1), jupyter(1)