Pies and Roses

On a slight tangent from my usual data, I always have the impression that there is more of a headwind when I'm cycling home, than when I go to work. Is it true?

Happily someone has kindly archived weather observations from the nearby Brussels airport, so I've been looking at the last 5 years' or so.

The question is, how to show it? For an R user, this is the fairly obvious answer - using coord_polar() in the lovely ggplot. To a trained eye, perhaps it's clear that going to work (heading North-East), typical winds are low whereas in the reverse direction the average wind is much stronger (the upper teens). 

But to me, the impact is back to front; it feels like it's easier, somehow, to cycle south-west. And there's no indication of how frequently the wind is in each direction.

So I want to reverse the scale. Now the length of the shape gives the strength of the wind, with the baseline on the outside, and the width of the base of the shape gives the frequency of the wind being in this direction.

Ok, so it looks like a messed up version of the dreaded pie chart, but is the impact of the message clearer?