Segmentation by understanding the differences between their transactional behaviours can be a way of treating customers in appropriate ways based upon their prior purchasing behaviour.
FastStats already has a number of different techniques for aggregating transactional values and returning a result for each customer record. These range from simple aggregations such as counting the number of transactions a customer has had, through to sophisticated pattern match aggregations as described in a previous blog (see 1).
However, to solve using the cyprus mobile number example existing aggregation techniques. This type of problems is exemplified where we need to work out the value of something for each product type. For example, which product have I bought the most of, or which channel have I mostly responded too. To solve those problems currently, we would need to create an aggregation for each type using a filter selection and then combine them using the function we need. As soon as we have more than a handful of product types this becomes too difficult.