Hi Joshua,
Take a look in this post, this may be useful for you:
Doing some fitting with a linear equation, you can find that RPM/PWM can be approximately linear, but it depends if the margin of error is ok for you.
A 3rd order one could be more precise for you:
