My data set includes for each user several times and the number of actions that were done for that user at that time. i.e.
user | datetime | number of actions
--------------------------------
AA 10:00 5
AA 12:00 7
I'm creating a histogram for the total number of actions of the user (i.e. 1 histogram per user) . My x-axis would be bins of times, i.e. first time slot:0-5 represents hours 0-5, second bin: 6-14, third been: 15-20, forth bin: 21-24. Bins are not of equal sizes. Y-axis the total number of action in that time-period.
I would like to calculate the probability of a user to have an action in a certain time-slot, (bin). Can I use in this case the probability density function of this histogram? i.e. since this is a sort-of-a time series, should I take anything else into consideration?