hellais
commented
2 years ago I guess the tricky thing with displaying percentages, is that while you are showing the percentage you lose the ability to see how many measurements contributed to that percentage (1 anomaly over 1 total is different than 10 anomalies over 10 total). How I generally address your this in charting is adding an extra axis to the chart, but it complicates the dataviz a bit
In any case I think we should be doing this type of calculation directly in the charting layer so we don't have to refetch the data for it and we can make toggling between the two views fairly quick
@agrabeli:
rather than replacing the anomaly + measurement counts with percentages, can't we just add the percentage as an additional column in the table? That way, users can see both the percentage, and the actual counts (and draw conclusions based on that).
If we are to add an extra column for percentages, the table is going to double in width, which I don’t think is going to be very usable. We could add it in the existing cells in parenthesis, but it’s going to make the table values much less legible.
We could however add percentages in the tooltips if you think that would be helpful
Yet, I mostly see the value of using percentages that of being able to normalise the data across time. I’m not so sure how helpful it is to look at percentages on an individual day
Community members mentioned that it would be helpful if MAT tables present the percentage of anomalies, so that they can quickly and easily evaluate if the number of anomalies is high in comparison to the total number of measurements (for the selected parameters).
Given that we already display the number of anomalies and the number of measurements in the MAT, it's a question of computing the percentage in the backend and displaying it in the MAT tables (generated along with charts).
I think this is very important feedback worth prioritizing.