Factoring in times would be even better! And perhaps other data such as the user’s age / average monthly spend etc. so that users could see merchants most visited by their peers?
I expect you’d want a pretty similar algorithm to the one that will be used to make the Monzo’s insights useful.