Solving the Mystery of Women's Sizing With Data Mining

Alternate Title: Yet More Things That Women Just Know That I Didn't Have a Clue About

I've sometimes asked my female friends what the sizes in women's clothing mean. i.e. what is a size 10, and what is the unit for ten such that a size 10 has ten of them?

Men's sizes are more straight forward - my pants are currently a 38/34. Meaning the waist is 38 inches around and the legs are 34 inches long. Men's pants are often in "relaxed", "regular" and "slim" fit, which seems to be a measure of how baggy they are. Given the same waist circumference and leg length, they can still range from "zoot suit" to "emo", but generally the main two numbers are still measuring something a tailor would recognize and correspond to on a tape measure. I have encountered vanity sizing for men's wear, but not often. 90% of the time, I can simply grab clothing off the rack and buy it knowing that it will fit correctly. No woman I know would do that unless they were very confident in the manufacturer, or buying a second piece of something they already knew fit.

Explanations for how sizing work are often vague and include equal amounts of guessing and exasperated sighs. The collected wisdom of the women I asked boiled down to this:

• Sizes are statistical averages.
  
• They vary by manufacturer.
  
• There is a high degree of vanity sizing in the system since there is nothing regulating what the sizes mean.

Complicating things is the simple truth that if you try to combine four numbers (bust, waist, hips, height) into one number (size), it's going to lose significance. In terms of computers, this is a form of lossy compression.

Then I discovered that a hacker in the UK decided to actually enumerate all the data and visualize it! She comes to several interesting conclusions, which I'm not going to repeat here. Go check it out, it's fascinating stuff.