It might seem a stupid question since I could not find anywhere a similar one, or maybe I'm asking something that defies laws of probability, but at least I'll know if this is the case.
Let's say that I have data regarding hotel bookings, and someone asks me: "What is the probability of at least 3 bookings being cancelled in the month of March?". Now, to answer that I would probably use a binomial distribution, however, to build one I first need the probability of a booking being cancelled in the month of March. How should I calculate it knowing that I have data from the past, let's say, three years? Should I just do a ratio of all the bookings cancelled in the month of March in the year 2018, 2019 and 2020 over the total bookings made in the same period? Or should I take the ratios of each month of each year and then do something with them?