There is an example strategy here that shows one way to do to the calibration, using a subroutine. Since you have a linear response, you don't need to use the more complicated scaling methods that the example shows, and you can stick with just a simple linear calibration.
That example is really more useful when the scaling is non-linear. For something as simple as this, an easy solution would be to set the input scale for the AIR40k module to 0-20000 ohms, and then in one of your charts create a loop (with a delay) that converts the input value to degrees. If you can orient the anemometer so that 0 ohm = 0 deg and 19999 ohm = 359.98 deg, it would be nice, but if there is an offset because north is in a direction different than the anemometer is expecting, you could include that as well.
(output in deg) = [(input in ohm) * (360 / 20000)] + (offset in deg)
Note that when using a offset, a check is needed to keep the output under 360 deg
if ((output in deg) > 360) then
(output in deg) = (output in deg) - 360
If it turns out that there is a scaling problem (0 is not the minimum resistance or 20000 is not the maximum resistance), you could include that too:
(output in deg) = [(input in ohm - minimum resistance) * (360 / (maximum resistance - minimum resistance)] + (offset in deg)