As a learning exercise, I'm working on putting together a few different machine learning projects that I can spend some time honing and studying.
One of these projects is predicting the locations of Uber surges, based on a third-party dataset of surge locations and times, and an assumption of reasonable seasonality to most of the factors that contribute to surge pricing. Essentially, I'm considering each surge as an "event" with an associated multiplier from 1.1 to 50 times normal price (the range of surge multipliers in the dataset), location as coordinates for latitude and longitude, and time as timestamp.
I'm fairly well-versed on models for working with time series data, but given the additional continuous two-dimensional location element, I'm a bit stuck on how to approach this, and I'm not aware of a name for this type of problem so I don't have terminology to search the literature for. Are there known methods for this?