I have a dataset that consists of a list of tuples:
(1, .5), (3, .7), (1,.1), (23,.8),...
In each tuple, the second element is the average price, and the first element is the number of items. So for example, (3, .7) means that three items were bought for an average price of .7.
From this list of data, I'd like to fit a power law tail (although really, I'm interested in the general version of fitting a distribution to partially aggregated data, in a nonparametric way). What are the common ways to split up, say (3, .7) into three separate observations, or more generally, how should I treat this partially aggregated data?
