/time_region_value: show the amount of missing data in the last year

[mindrones]

The data collected in the last year of a series could be incomplete for a number of reasons, for example:

• we collected data in the last year of the collected data series, so the data is likely incomplete
• we collected the data after the last year of the data series, but the provided data for the last year correspond to a date in the mid of that year

The consequence of this is often seen in trends, with the last data point being at odds with the general trend [1], which might trick the user into thinking that something happened in that year, or it might seem an error. Moreover, in the map it would be impossible to see this, leading to possible misinterpretations. So we need a visual representation of this, not necessarily in the map or in the trend charts.

We could have the last circle in the timeline being a mini pie with a black sector representing how much data we are missing, easily doable by using data_date: e.g. if we miss 5 months worth of data, we’d see a black sector with an angle of 5/12th of 2π, 150º.

Note that data_date is currently optional so we would have to ignore the problem for those indicators where data_date is not provided. Alternatively we could make data_date mandatory, TBD.

Other possible but less desirable solutions could be:

• use a tooltip on the last circle of the timeline (might be annoying after a while, and it would be visible only interacting with that last year, probably when the user wants to click on it);
• explain this in the app guidelines (which user might not read);
• explain this in the interactive tutor (which which we should provide later on, but that has the drawback that when the tutor is deactivated we can't be informed of the problem all the time).

[1] data_date: 2021/07/01

[georgerichardson]

I like the idea of visually indicating it. However I'm not sure it will be a consistent effect. Let's say that there is an indicator that relies on a statistic that is only published once per year during the summer, then the data will not be complete. There will only be incomplete data if it relies on some underlying continuously published stat.