Self-learning platform

One of Datagrok's unique features is that most data operations happen within the platform. Beyond the convenience of a single tool with centralized access management, this lets the system learn from observed user behavior.

Think about Netflix's movie recommendation engine, but instead of dealing with just two entities ( users and movies) and one relation (user's score for the movie) we have a much more complex case. We got dozen of different entity types (such as query, viewer, etc), connected with different relations (such as 'query ran_by user') and restricted by different constraints.

When enabled, the self-learning component uses AI to spot usage patterns and suggest relevant actions. For example, it might suggest a visualization for your dataset, apply a prediction model trained by a colleague, or create a derived column like BMI from weight and height columns. Suggestions draw from all users' activity, not just yours, helping spread organizational knowledge across departments and time zones.