Well, start by deciding your limits of measurement...

For the sake of argument, let's over-simplify this...

If I was doing wind-speed down here on Earth, then I'd choose limits between 0 and say 150mph.

Next I'd decide how I'd measure the Wind-Speed... So I'm going to have a Vane diameter of 200mm... That means one rotation equates to 628mm of travel...

150mph equates to 67 metres per second...

67 metres per second divides down to about 107 pulses per second.

Is this good enough for what I want? If I was doing it, then I'd probably say No - mainly because I'd want more pulses per second so I could improve the resolution (unless you are going to have a really long sampling time which then means you won't be able to measure gust). And so the design process continues.

Only once I know what the worst-case scenario is, would I then decide on how my PIC is going to process this information - that's the easy part!

You want to run with a 20MHz PIC... probably total overkill for this application.

I certainly wouldn't go with a shaft with a contact on it (because contact bounce will give you nightmares)... I'd consider a magnet bonded to the shaft with a Hall-effect switch, or a slotted vane and an optical sensor (best option for improving resolution at minimal cost whilst still keeping the shaft balanced for reduced frictional losses).