Sure; so lets say I wanted to predict the price for a rune scimitar. All I would need would be the prices of the rune scimitar at different periods of time. By looking at the trends I would try to fit different forecasting models to predict future prices and assess the fits and assumptions of the models. The actual fitting of the models can be done using the programming language R which can read in data in a table form and do all the computation. I would also take in to consideration the effects of different patches, however I would worry about this later.
If you're interested you can read more about the methods I will be using here
https://en.wikipedia.org/wiki/Forecasting under Time Series Methods.
Another thing I could do would be to determine if there is a relationship between any 2 (or more) items. For example, perhaps when the price of a rune scimitar increases, the price of lobsters go down (maybe people will not be able to afford rune scimitars and thus pk less and not require lobsters). Determining this sort of relationship is generally much easier than forecasting.