-
It is only important - in this case - that the data is collected consistently day to day. As long they stick to the same methodology while building that chart, it is an accurate representation.
It isn't the numbers, but the comparison between the numbers that matters.
beseku-
Well not quite. For example, the number of tests run is pretty important too. We've seen it in other places before, a drop in the number of new infected one day, correlating with a drop in tests performed that day.
Also, that doesn't change anything about the fact that the data is pretty noisy, and basing anything on single-day numbers is a bit shaky. See the FT graph above, where they add up all new cases of the last week on each day, which makes for a much more convincing measure.
SwissChap
How are they testing in Spain though, and how much? I'd be very careful with getting any conclusions at all from the case numbers - I think, if anything the death numbers in the second table are more 'solid' in that sense. Also, they don't mention basing their calculation on some kind of running average (unless I missed it?), so while the decrease in speed is of course great to see, the numbers would be a lot more dependable if they weren't exposed to the types of random ups and downs (noise) in the daily data.