Premise 1: We measure how many people enter an ice cream store at any given hour during the day, for the entire year. The graph will look like a very long zig-zag pattern of 8760 hours/ticks.
Premise 2: We measure the same thing for a fast food restaurant. So we have 2 arrays/sets of amount of people entering the respective building at every given hour during the year.
Premise 3: Now we plot the distribution(graph of "How many times do X number of people enter the building at any given hour during the year?") of each, and find that they are nearly identical in kurtosis and skewness. (height, width and position)
Premise 4: One day, we lose the measurements/data of March from the ice cream store.
Question: Can we copy the measurements of March from the fast food restaurant to fill in the missing space of the ice cream's March-section? Would this be mathematically valid/justified? Can we safely assume that the March will be "close enough" of an imitation of the original so that we can work with the data?