How do you ask for a data dictionary & expectations from the data owner?

Originally posted in slack here:

If you’re an engineer or project manager but you’re not the data owner, how do you get that information from the data owner? Please share what you’re currently doing if you’re in the same situation!

For us, the data owner is usually not us and they are also usually not technical so sending over a jupyter notebook can be confusing and too much of a hassle.

As a start, I created a google doc with just a table that could be filled out but there’s still plenty of room for improvement. Someone suggested using more natural language for expectations instead of the expectation name (e.g. expect_column_values_to_not_be_null(school_code) would become ‘School code should not be empty’). A form might also work better or a google sheet. Ideally we would be able to also go back to this document as “sign-off” from the data owner in case there were any data issues down the road.

1 Like

@porterhaus this is becoming a pretty common situation. We are working on a SaaS product that will complement the open source Great Expectations and will streamline the process of working with “non-coding” stakeholders.