Len Kiefer, Deputy Chief Economist at Freddie Mac, recently posted an analysis of global housing price trends based on the international house price database (from the Dallas Fed). Using those data, Kiefer compared housing trends price increases (and in a couple of places like Spain and Ireland, decreases) across 24 countries. He also looks at the change in affordability of houses, comparing housing increases with disposable income growth. (Belgium is a surprising outlier in this category, where housing cost has inreased 1.8 times faster than income.) But I thought the most interesting chart was the animation of house prices below: despite the much-publicised spike in housing prices in big cities like London and San Francisco, when you look at the country-level data it's actually Canada, Australia and New Zealand that lead the pack.
Len used R for all the analysis and visualization, and helpfully posted the R code to the end of his blog post. Kiefer imported the Excel data file from the Dallas Fed simply by using the read.xls function from the gdata package. The charts were created using the ggplot2 package, and the chart above was animated from individual frames using the animation package.
(Incidentally, there's a new way to create animations using ggplot2: the gganimate package. This PlotCon presentation by David Robinson explains how to use gganimate, including the use of the tweenr package to include smooth inter-frame transitions in the style of GapMinder.)
For more details on the housing price analysis, including the R code, follow the link to Len Kiefer's blog post below.
Len Kiefer: Global house price trends (via Urban Demographics)
As a former statistician and economist in a research dept of a central bank I love these types of analysis. While it is great use of gganimate there are a host of unwritten assumptions underlying this analysis. One of them being the prices are not corrected for size of asset (i.e. non hedonic case shiller) rendering the comparisons of countries very misleading. Even within country it is very problematic since the variance is usually high. So any conclusions regarding outliers should be made cautiously.
Posted by: Yoni Sidi | December 12, 2016 at 17:15
moreover: I wonder what the rural/urban split is in Canada, Australia, and New Zealand vis-a-vis the USofA (the instigator of the Great Global Recession). IIRC, all three are rural wastelands sprinkled with a few cities whereas the USofA has a larger proportionate rural population (we got a lunatic president thanks to them), and cities are where house prices driven up by demand forces of increasing urbanization.
Posted by: Buggy Funbunny | December 13, 2016 at 04:59