GIVEN: A company has multiple "retail" locations (10 as an example). They collect data on customers, so they are able to identify customers that shop only 1 location versus customers that shop at multiple locations.
They have a set of data that shows how many customers are shared per site, and how many out of those customers are shared with each additional site.
For example:
Store 1 has 12007, of which 6234 also shop store 2, of which 447 shop store 3 and so on.
Store 2 has 10397, of which 6234 shop store 1, of which 459 shop store 3 and so on.
Here is a grid representing my dataset. The "shared with" columns are shared out of the "shared customers" column, for each row.
To be clear the "shared customers" column contains NO customers that ONLY shop the given location. That means that every one of the 12007 customers from store 1 HAVE in fact shopped one or more of the other stores, and the number of customers out of the 12007 is shown in the "shared with" columns.
PROBLEM: Based on this set of data, how can I count the number of unique customers within the set? If the company only had 2 locations then this would be easy. Count all from store 1, add store 2 and subtract the number they have in common. However, once you add a third location, I am not sure how to count them, or even if it is possible.
It is not possible. Suppose you have five stores, each store with six shared customers and sharing three with each other store. That could come from one customer shared with each other store and two customers who shop all the stores, or from one customer who shops each set of three stores.