I have a network of 100 bank accounts at 10 banks and I want to analyze the flow of money between these accounts with a weighed social network analysis. However, I can only see the balances of these bank accounts (not the transactions) on four dates from the start: day 5, 10, 20, and 60. I also want to incorporate demographics in my analyses--does age, income level, occupation, etc. affect an account holder's likelihood of giving or receiving money to another account holder?
I am doing my analyses in R code.
- Is a social network analysis appropriate for my questions? If not, what other models or tests should I try? If you are familiar with R, which package(s) would you recommend?
- How do I weight the edges if I don't have discrete transaction values between accounts? (i.e., I can't measure how much money was moved between two accounts, I can only look at the balances of those accounts on the same day multiple times.)
[I'm using bank accounts as a metaphor here because my data are complex chemical values (change of the change of the ratio between two molecules; stable isotope tracing) in an ecological community. Each bank account is an individual plant and the balances are the measures of chemicals. The details are unimportant to my conceptual question, though.]