I'm trying to model the age of inventory at time of sale, where a given sale (say, for \$10,000 worth of a particular product) might be fulfilled with \$5,000 of stock that is 1 month old, \$3,000 worth of stock that is 3 months old and \$2,000 worth of 7 month old stock. I have data for 1,000 or so products with age of stock ranging from 0 to 18 months. If I didn't have to include the amounts in the analysis, it's a straightforward negative binomial modeling problem, but I'm not understanding how I can weight the events by the amounts.
Here's what the (weighted) data looks like:
Any ideas?
