Redacted Application Submissions

[apricot13]

On the DPR we have a requirement to show the original application submission example here: https://planningregister.org/camden/25-00114-HAPP/application-form

Our implementation was done without access to this schema so redaction is done by us manually excluding certain fields, but we also didn't have this schema to refer to for potential fields so to be on the safe side we're currently not showing all the data.

We'd like to propose a redacted version of the application submission schema. Suggestions for an approach to this would be

• separate versions of each schema minus the redacted fields
• additional annotation of the fields to be redacted @redacted
• additional base interfaces with the public content and redacted content extended from them

For the responses I believe we can add a 'sensitive' flag into the metadata which can be manually reviewed in the back office system with a permanent 'sensitive' flag applied to certain fields somehow to reduce human error?

[apricot13]

Meeting update Tuesday 8 April 2025

Some further context on the redaction:

The diagram shows how all the schemas connect with one another

Areas that would require redaction:

• Responses in the PrototypeSchema (Application submission)
  • contains the user submitted data with the question asked at the time alongside the metadata
• Data in the PrototypeSchema (Application submission)
  • the Responses section transformed 🪄 into structured data would require various fields to be redacted
• Certain fields added by PublishedSchema eg PlanningOfficerRecommendation
• Documents in PublishedSchema
• Comments in PublishedSchema

Notes

• Everything but the Responses section can (hopefully) be programatically marked as redacted using typescript/json-schema
• The Data is derived from Responses so would displaying just Data suffice?
  • the Responses provide the questions as context for the answer and the extracted data so would including Responses[].question and Responses[].metadata.date suffice?
• The comments process is pretty much there for redaction
• The process will rely a lot on people

Questions

• how to mark the types as redacted
• how to mark the json schema as redacted
• do we need to consider including routes (API and processes) for reporting missed redaction's
  • wider considerations around abuse of this process

Actions

• Update the Comments and Documents PostSubmission schema to include redacted url and redacted comment ready for PublishedSchema to exclude/remove those

[tom-gayler]

BOPS approach to handling sensitive data on application form https://docs.google.com/document/d/1oCH8QgtEmbz02q5u7MskRcuwOyfSFcaLuuznM-9AuPw/edit?usp=sharing