Calculate grades

[waldoj]

Every state and every dataset should be graded.

[waldoj]

The simplest approach may be to calculate grades within Google Sheets. That way they're stored in a canonical location. Downside: the formula isn't included within the site source. Solution: Put it in the repo's README.md.

[waldoj]

I think that we need to weight machine readable far higher than anything else, because that's the crux of this census. This is the difference between data and not-data.

[waldoj]

So if this is a 100-point scale, and we've got 12 metrics, then we can do the math to divide them up. Imagine that we give every metric the same score, except for machine readable, which we score far higher. (Complicating things, some metrics imply the responses to others, e.g., if any field is yes, then exists must be yes. But let's set that aside for the time being.)

I think we can ignore exists for scoring purposes. I think that machine readable should be worth ~half of the entire score. So that gives us 10 metrics plus machine readable. So that suggests that machine readable should be worth 50 points, with the remaining criteria each worth 5 points. That yields a maximum possible score of 100 and a minimum score of 0.

The only wrinkle is that if a dataset exists, but it isn't public and we don't know anything about it, it gets a score of 0 instead of a score of, say, 5. But in a census of open data, perhaps a 0 is appropriate.

[waldoj]

I've got scores now, but I'm struggling to calculate grades within Google Sheets. (It's been years since I've used Excel formulas, particularly the LOOKUP function.)

[waldoj]

Victory!