The populations they used in their calculations are a bit wonky. For example, Ottawa apparently lost population from 2010 to 2011 according to the chart, and there's obviously something off with the per capita power centre number for Toronto. I reran the per capita figures using the
actual Statistics Canada population estimates for CMAs, and the picture is roughly the same as here except for a few details. The general per capita figures for shopping centres and power centres is slightly lower and Ottawa's increase in per capita amounts is less than shown. One key difference is that Winnipeg had a per capita increase in shopping Centre and power centre s.f. per capita instead of a decrease, Ottawa had a decrease in power centre s.f. per capita, and Toronto's power centre s.f. per capita only increased slightly instead of by 50% or so.
I don't normally try to replicate results that I come across like this, but the numbers seemed off to me. It makes me realize just how often I take graphs and charts like this for granted, though. Perhaps I should do this more often.