The scoring system on the datacard consists of two main components: the Datacard Score and the Freshness Score. The Datacard Score is derived from two critical aspects:

  • Characteristics - accounts for 80% of the score, emphasising the dataset's relevance and applicability for AI models.

  • Licenses - contributes 20%, assessing the dataset's legal usability and openness.

This weighted approach ensures a balanced evaluation of both the dataset's content quality and its compatibility for use in AI projects.

Last updated