Here again are the charts for the active cases and case trend in Minnesota. The original analysis was a bit complicated and tried to use the state’s formula for no longer needing isolation compared to cases reported by date of test. Probably pretty accurate. The other two charts on the comparison of methods chart use a flat 10 day rolloff and a flat 7 day rolloff. The ten day looks close to the original method because the state is basically using a ten day rolloff. The seven day may be more reflective of the real infectious period, and because of the faster rolloff creates a lower peak. Thanksgiving is seen as a bump down because of so little testing in that time period. That’s right, the virus knew not to infect any one over those days, so no one thought they needed a test. Anyway, it is very clear that mitigation measures have little to do with the shape of these curves and that active cases peaked before any governmental orders. It is likely that voluntary behavior changes in advance of those orders may have affected the peak a bit, but if you smoothed out the sawtooth weekend effect, you wouldn’t really see any noticeable difference from a classic epidemic curve. On the raw number change and percent change, both based on the original analysis, that spike you see in early December is caused by the extremely low testing and cases during Thanksgiving, which rolls through 10 days later, appearing to cause a spike in active cases, but it was just an artifact from so few cases rolling off. Thanks again to DD for doing the work on this.