Around the time of the Renaissance in Italy and Germany artists like Albert Durer would spend a considerable amount of time coming up with the ideal proportions of the human figure, including the human face  (see below); it was all based on the Golden Mean.  Durer’s manuals on The Four Books on Human Proportion are particularly relevant here, and come to my mind.

Construction of a golden rectangle: 1. Construct a unit square (red). 2. Draw a line from the midpoint of one side to an opposite corner. 3. Use that line as the radius to draw an arc that defines the long dimension of the rectangle.

Today, an article in the New York Times comes us with the Modern Equivalent of Durer’s books on Human Measurement – The Sum of Your Facial Parts

where the human face, in a photograph, is subtly altered to fall within a standard set up beforehand.  The results are very hard to notice 80% of the time, but somewhat noticeable the other 20% of the time.

“…. The software program, developed by computer scientists in Israel, is based on the responses of 68 men and women, age 25 to 40, from Israel and Germany, who viewed photographs of white male and female faces and picked the most attractive ones.

Scientists took the data and applied an algorithm involving 234 measurements between facial features, including the distances between lips and chin, the forehead and the eyes, or between the eyes.

Essentially, they trained a computer to determine, for each individual face, the most attractive set of distances and then choose the ideal closest to the original face. Unlike other research with formulas for facial attractiveness, this program does not produce one ideal for a feature, say a certain eye width or chin length.

The question I have – does it really matter?   As an analyst, I do care about subtle things – in fact, it’s those connections that make delving into data to find what was at once, hidden, and revealing it, that makes the job of Web Analytics interesting.

But in the case of using a process to improve “imperfection” one whas to wonder.  For example, there are cultural differences and a small number of individuals that came up with the formulas”

“….68 men and women, age 25 to 40, from Israel and Germany, who viewed photographs of white male and female faces and picked the most attractive ones.”

I don’t think that’s enough people – 2000, maybe – 6000 – 10000 would be better – but the next question comes up – if you had the extra 9,932 “panalists” data to draw on, would the formulas be any different?   To be honest, probably not that much – with a big if …. depending on how well the initial group of 68 people where choosen.

Recently I’ve been working on a Paid Search Project – and noticed just how much of the online spend and converstions are focused on terms my client has in the first position, and usually, on Branded terms – the “Long Tail” might alter the results somewhat – but one wonders if it is worth the effort.

Same here – I believe the 68 panelists whose ideas of beauty were assembled to create the program, is too little – but then, Albert Durer, back some 500 years ago, had only one person to decide what his alogrythm was – and it worked … at least, for him.