I think it depends a lot on what you want to do.

For any kind of machine learning or data processing, linear algebra is definitely a good thing to have. Basic calculus is good too, but I'd say may be less important if you're applying methods and approaches as opposed to developing them too. Differential equations, while providing an important set of mathematic tools which may open some doors you didn't even know existed, may not be the "highest yield" material from a programmers point of view.

In my opinion, probability and statistics is a big yes for anything data related, doubly so for algorithms. I'd also consider some graph theory, as it's amazing how many problems can be projected as networks and then approached from that perspective.