First a global view of the dataset :
About 800 controls per week, except during the lock-down in 2020, and a slightly lower control pressure in 2021 and 2022.
Poor results (To be improved or To be corrected urgently) are stable at around 6 %, except a recent spike in 2023?
It seems that poor results increase during the year, from June to November.
This surprising periodic phenomenon is also visible by day :
So for the « when », it is : not in summer or autumn.
What about the « where »? It seems you also could be careful in some départements…
Not good in Guadeloupe, Guyane, Réunion and the southern lower Seine valley, west of Paris.
Can we see more in details ? Using a 30 km kernel smoothing :
It confirms some hot-spots west of Paris, in Alsace, in Indre, Cher, Alpes-Maritimes and between Gironde and Landes. You are safer in Paris and Bretagne…
Has it changed ?
No real trend…
Download the supporting data and R scripts :