Screen Shot 2020-06-29 at 11.10.30 PM.png (89.32 kB . 1226x622 - viewed 754 times)
This data (screen grabbed from the New York Times just now) clearly shows a periodic (possibly sinusoidal) variation in new cases reported each day. The period appears to be is exactly 7 days (1 week). With maxima coming on consecutive Fridays (with the occasional Thursday).
I see a few possible explanations:
1) This is an artifact caused by weekly patterns in data reporting (file the report by the end of the week.)
2) This is an artifact caused by when it is easiest for most people to get the tests.
3) This is real and represents a weekly pattern in when people get sick (maybe folks go out Saturday night, get infected then, and then feel bad enough to get tested 6 or maybe 5 days later.
4) This is real and represents ripple effects of more people get infected when there are more infectious people around, and because of the induction period there are actually bulges that propagate at a period of about 1 week.
5) something else?
Looking at the deaths graph offers more insight:
Screen Shot 2020-06-29 at 11.15.34 PM.png (113.1 kB . 1264x678 - viewed 752 times)
1) is challenged by the fact that the reported deaths, although also following a weekly periodic undulation, peak on Wednesdays. (presumably there would be similar reporting biases)
2) is also not supported by the deaths graph (people don't tend to only die when it is convenient)
3) could be supported, if the length of time from showing symptoms to death is fairly consistent (and not supported if it varies widely)
4) see above