One also could claim the protests are the cause of these upticks. I don't, because the data aren't there, but it is one possibility.
I think the lag is even beyond 10-14 days. Remember that exponentials look small at first, and when R0 starts getting above 1, it's an exponential.
Let's say 10 protesters at a 10,000 person protest had COVID-19. Each of those protesters infects 10 people--which might be high or low in that sort of environment. That's 100 people.
Well, a week or even later (considering a several-day lag between an administered test and getting results) those 100 people infected by 10 will be barely a blip in the radar.
Now let's say that right before those 100 people started showing symptoms, they've infected 3 people each before quarantining. Given that they were willing to go out in public protesting, maybe they're not exactly social distancing well anyway. So now you get 300 more cases, but again it takes another week+ for those 300 to show up. You barely start to see a turn in the trendline.
When those 300 infect another 900, and when those 900 infect another 2700, then you start seeing the numbers turn in big enough ways to really notice. But now you're a good 4 weeks post-protest before you've seen the numbers reflect the spread that occurred at those protests.
We're impatient. If we want to blame protests we think we should see 2700 cases pop up within 7 days of protests starting... That's not exactly how exponential curves of infectious diseases look.