In theory it should. Basically, when you start sampling a specific type of item it creates spectral charts. They anomalies in the charts that distinguish zinc from copper from nickel may be small so you may have to zoom in to observe them, but they should be there and you should be able to train the system to understand the difference.
I personally have only used existing apps with the device, I haven't tried coding anything yet with it or analyzing the data that comes from it. I would encourage you to check them out and ask them if you are that interested: