In order to have deeper insights into the distributions of delays, would it be possible to try:
removing the long right tail that is likely to be errors again;
also considering actual number of shots instead of frequencies only so that we can make comparisons between months.
As for the month to consider, I find the the month of expected vaccine shot much more meaningful that either the month of birth or the month of actual shot.
In order to have deeper insights into the distributions of delays, would it be possible to try:
As for the month to consider, I find the the month of expected vaccine shot much more meaningful that either the month of birth or the month of actual shot.