Document the Forecast data model

[jack89roberts]

Couple of things I noticed so far:

• Placeholders are treated differently to users in the data, e.g. they have a "placeholder_id", not a "user_id".

• Some project names appear multiple times, with each instance belonging to a different "client", e.g. Safe Haven (SPF), and Safe Haven (Tools, Practices, and Systems). In the projects table data extracted from Forecast these then appear as two distinct projects with different ids. I don't think this is ideal; they should rather appear under one project id with multiple client ids.

Safe Haven in Forecast interface:

Safe Haven in Forecast projects data:

• I'm struggling to find data on the total required resources for a project, e.g. "project x needs 2.0 FTE from April to June" type data. Strangely, I don't think this exists in the forecast data. The harvest data has a "budget" entry, which I think would be the right place, but very few projects have a budget entered. I'm also not sure this gives the flexibility to have changes in budgeted FTE during the project.

Example project data from Harvest:

[jack89roberts]

• I don't think the Forecast data has a flag to explicitly say whether someone is currently active in the REG group or not (e.g. Angus is in the data). The Harvest data has an "is_active" flag.

[jack89roberts]

Weekly capacity isn't populated for all users in the data - seems like if the capacity is 40 hours/week (default) it's missing in the data, but other values make it into the data.

[jack89roberts]

What does archiving someone do, and is that the right thing to do for people that have left?

[jack89roberts]

See: https://github.com/alan-turing-institute/WimbledonPlanner/blob/master/notes/Forecast_Data_Model.md

[martintoreilly]

@jack89roberts Cool data model pic. Did you make it yourself? If so, what software did you use?

[jack89roberts]

@martintoreilly I used this and then inkscape to play around with a couple of things: https://dbdiagram.io/d :)