I'm surprised how good that linear fit is, considering all the factors that could skew it. HDD indicates how long * severe a winter is, but insulation thickness in a passivhaus with adequate thermal mass really only depends on the coldest week, and sunlight levels.
I see you have R=0.90. Can you get that higher by correcting the Moscow data point? (I assume that Moscow is colder than Brittany, else why would Napolean have run into trouble?)
|