(Click on figure for larger image and use browser Back Button to return to blog.)
Hat tip to Gregor Aisch, Adam Pearce, and Steve Hoey, and sourced from the mashup dataset and visuals by Lisa Charlotte Rost.
Mr Aisch’s innovation was to use Loess regression to display. Loess is one of a set of local regression methods. I personally prefer p-spline smoothing (penalized spline regression).
In contrast, Boston is 42.4N, Bangor, Maine is 44.8N, and Montreal, Quebec, Canada is 45.5N.
See also article in Quartz.