Data modeling

[mgwalker]

ID the exact data that analysts need to use, and that policy dictates

[mgwalker]

Based on some early reading (mostly looking at detailed budget tables we got from states, and then using that to inform a bunch of Google searches), here's my initial impression of the financial data that we'll need.

Each line item will need three pieces of information: federal fiscal year (FFY), total cost, and type of cost. We can combine the total cost and type of cost to compute the federal and state shares of each line. There seem to be 7 major cost types, with two of them split into 6 more:

• 2A: Cost of in-house design, development, and implementation (DDI)

  • 90% FFP for planning, RFP work, requirements analysis, direct costs, testing
  • 75% FFP for "site preparation"
  • 50% FFP for, I guess, everything else? (specific examples are feasibility studies, training DDI staff, and indirect non-personnel costs)

• 2B: Cost of vendor design, development, and implementation

  • 90% FFP for planning, RFP work, requirements analysis, direct costs, testing
  • 75% FFP for "site preparation"
  • 50% FFP for, I guess, everything else? (specific examples are feasibility studies, training DDI staff, and indirect non-personnel costs)

• 4A: Cost of in-house operations, 75% FFP

• 4B: Cost of vendor operations, 75% FFP

• 5A: Cost of non-approved mechanized systems, in-house, 50% FFP

• 5B: Cost of non-approved mechanized systems, vendor, 50% FFP

• 5C: Cost of non-approved mechanized systems, interagency, 50% FFP

[mgwalker]

At this first look, it seems like we only need 11 numbers per federal fiscal year (the total for each FFP level of each type of cost). From there, we can derive/compute/copy the 39 numbers that are in the MMIS MDBT template.