Someone messaged me and said that using 7 day buckets for the data can be manipulated, so I'm doing another version of the graph with a 4 day moving average to split the difference between the single day source data and a weekly view that I wanted to look at.
In my original comparison we entered phase one at 30 new cases per week and exited phase one at 60 cases per week. His complaint was that is cherry picking since shifting the numbers by 2 days makes that comparison 26 vs 31.
The moving average was 6 at the start of the phase 1 reopening and was at 9.75 at the end of the phase 1 reopening. Not double, but clearly an increase.
View attachment 546748
I'm not so sure about that. With a lot of ups and down, there are always problems one way or the other.
If I look at the 4 day moving average, and do an additional interpolation in my mind, I'd think the first value should be around 3.5, and the second around 7. (Although the second depends on how it will continue.)
That's still double.
The main problem is that you have an exceptional upswing at the end, and not enough data to tell if that's an outlier or the beginning of a real trend.
I haven't used 7 day buckets very much yet, but they seem promising to me. If you have the option and patience (or software), it might be worth trying to move the cut off points between the buckets. But in this case you can tell it is still difficult to make a judgement because the 7 day buckets still have a lot of up and down. So you might think of using 14-day buckets in this case, starting at the end going backwards. That will probably show a much more even picture, but not answer your question of what to make of the upswing at the end.
However you do it, I think the problem is in this case you need future data to evaluate the upswing-or-not at the end.
EDIT: If you take your 7 day buckets and combine them into 14 days buckets, starting at the end, you get this sequence:
48, 47, 24, 30.5, 42.5.
This is much smoother, but if you move the cut off points, it probably wouldn't go as low as 24. That may or may not be better. It still suggests a strong incline in the last 6 weeks