The general logic is that when fewer people are in contact with each other, say at stores/pubs/restaurants, work, transit stations, the number of new cases drops. In this chart below, you can see new cases with respect to mobility data: transit stations, retail/recreation (pubs/restaurants/movie theatres), and work. The dips (in orange) correspond to when people were at home and not working (Memorial day, Thanksgiving, etc.). The spikes (in black) correspond to people going out to shops/pubs/restaurants; for example the tall black spike (on right) is "black friday" in the US, a major shopping day for retailers. The grey "shadow spikes" is new cases, and you can see a rough correspondence between black spikes (or sustained elevation of black), when people were out and about, and a few days later an increase in the grey "spikes."
View attachment 149123