The curve shows the daily number of cases, the dashed green line shows the start of the curfew in Paris and 8 other cities, the dashed purple line the extension of that curfew to 54 departments and the orange dashed line the start of the lockdown. 
It's pretty clear that incidence started to fall before the lockdown was implemented, so while it may have accelerated the process, it would most likely have happened without it. It's even clearer when you look at what happened in Paris. The curfew may have played a role though.
The curfew made it illegal to leave your home between 9pm and 6am, but if it was responsible for the fall in incidence (which in my opinion it likely was at least to some extent), it's probably because bars and restaurants were closed, not because people couldn't go out per se.
Sorry, I was actually double-counting so the number of cases was way too high, but it obviously doesn't affect my remarks about the effect of interventions since it doesn't change the shape of the curve. Thanks to @houndcl for catching this! Here are the correct charts.
• • •
Missing some Tweet in this thread? You can try to
force a refresh
