As an SBA user, I need to know the final form and content of the post-joined data being submitted by the agency to Treasury

[bsweger]

Acceptance criteria:

• [x] Answer to the question: how many files will there be and what do those files represent?
• [x] Determine format of files being submitted via website (.csv, etc.)
• [x] Document the data elements included in each submission file (or single file), depending on the answer to the 1st criteria

[bsweger]

Based on the schema work that Treasury is doing, we've decided to proceed with four input files for this project. Each file represents a specific level of aggregation; this work is in progress here: https://github.com/18F/data-act-pilot/pull/107

Got some great feedback from Kavya today. We reviewed the four files, and she had these comments:

• The source column should reflect the reality of this pilot and this current sprint, not our best guess at what the final spec will say. For example, since we're pulling recipient info from Prism and not from SAM right now, the files should note "award system" as the source.
• Same for any accompanying diagrams.

She did think that splitting the data out as specified in the four files is do-able from the SBA perspective.