Re: 'Data Uploader'

[vonbraunbates]

Problem with 'Data Uploader' (https://user-guidance.analytical-platform.service.justice.gov.uk/tools/data-uploader/)

Feedback from external users of the uploader:

• They cannot view the user guidance or the repo README. Even though the steps are fairly self-explanatory, it would be helpful to have a "boilerplate" text to email to them so they know what to expect.
• To create a new database, users need both a data movement form and an asset registration form filed on OneTrust.
• There is a missing underscore at the end of the prefix (so ...uploader_prod_ not 'uploader_prod) which was causing users to start their database name with an underscore.
• Good database names:
  • should say something about the source (e.g. procurer or system).
  • There's not really enough space left so describe the policy domain.
• Good table names:
  • should include a domain,
  • be descriptive,
  • avoid acronyms (unless in the acronyms repo ),
  • and flag sensitivities around access (e.g. if there's personal data in the table).
• Once the data are uploaded, someone needs to make it available via the database access repo
  • Internal users can flag this on Slack in #ask-data-engineering
  • Sensitive data will need the email of a required approver to sign off approvals
  • Once the new data are added to the sources in the access repo, users can add a project access request to access it from the AP.

I think it is worth specifying guidance around database and table names, since I have seen things like pv_data.data and rctl.leads which makes it difficult for data engineers to process access requests. (The MOJ Tech Guidance talks about naming services, which is very different to naming data.)