Indicator coverage visualisation: calculation

[pindec]

The indicator coverage calculation should follow these existing resources:

• OC4IDS use cases #shared
• OC4IDS indicator coverage notebook

Camila is updating them in preparation for the sprint, both in terms of checking the fields and calculations, and looking at rationalising the topics (see CRM #5569).

We should only calculate indicator coverage for indicators whose questions have "Yes" in the "Possible with OC4IDS?" column in the 'Use cases' sheet.

For each indicator, we should calculate:

• whether it is measurable (yes/no) - i.e. if the fields required to calculate it according to method_1 are present anywhere in the data.
• the indicator coverage metric shows the percentage of projects in which all the required fields occur.
• the "missing fields" where indicator coverage < 100%, and corresponding coverage metric, which shows the percentage of projects in which all fields are missing (i.e. users can use this stat to target which fields to gather)

Limiting the calculation to method 1 only at this point will simplify the calculation and interpretability (since we will not need to explain why there are multiple methods). There is an implicit hierarchy in the method ordering.

[jpmckinney]

As in #102 (comment), have all the percentages mean the same thing. That is: a high % means high coverage. Do not do "percentage of projects in which all fields are missing".