I have a data set with payment instances. User pay an online retailer through a system. The system captures user payment instance (e.g. how much they pay, when they pay, what method they use to pay, who is the retailer, ...). This instances are unique to each users, that means you have many of these instance for a particular user.
I want to create a meaningful category of dataset from these instances or events to describe a person. It is like creating a survey question from these instances to prepare it for statistical analysis to understand user payment behavior.
Anyone did this before or any suggestion how to go about it?
Not sure what the goal of your research is, but it seems to me you want to use a multilevel growth model. The system will have logged the date of purchase I assume. With this you can model a time-dependent (growth?) curve. I suppose you want to model changes in payment and not have an overall indication of paying.
Otherwise you can aggregate data (for instance total amount spent and number of purchases (to calculate average spending as total amount spent/number of purchases). Most characteristics of persons change over time as well, like wealth, income and age), a few remain unchanged (like gender, but a small amount also change 'gender'). What you can do depends on the background variables you have access to.
Maybe you can provide is with more information, so we can give you better advice.
Thanks Peter Moorer for your answer.
I am working on classifying users' financial well being (FWB) state. In the end the aim is to create a FWB scorecard for each users based on their payment behavior. (e.g. they are healthy, just feeling conformable, getting by ...).
The data set I have is series of instance of their payment with additional information around it such as erratically spending habit around salary or others. Image it as a paypal payment data.
So my thought is to make some 'insight' from patterns i am looking at the data. Since it is an instance of payments, i 'should make' into a survey like variable. That is the challenge i am having.
- Badges
- Science topic
- Similar topics
- Medicine
- Physiology
- Biosignals
- Biomedical Signal Processing
- Statistical Analysis