Growing out of popular culture

Comparing PageRank tendencies of 'liked' pages by age.


Topic & data gathering

Thinking about my network on Facebook, I made the following hypothesis:
older people are more likely to 'like' mainstream websites than young people.

To test hypothesis (at least within my own network) I used the Facebook API to find out which websites my Facebook friends liked. A simple Processing script looked up the Google PageRank for every website, giving a good indication of its popularity.

All this data I put into a spreadsheet, grouping it by age. To my surprise, on average, older people in my network actually liked less popular websites! Now how to visualize this data meaningfully?

1st version: scales

In the first iteration, I tried to include a lot of information. The visualization showed each individual as a circle, its size and color indicating the average PageRank and amount of likes the average was based on.

I used the visual metaphor of scales to show that the data was being compared. Unfortunately, this was misleadning, as the number on the scale indicated an average, but the size of the stack was a sum.

2nd Version: blobs

The next iteration did away with the misleading scale metaphor, but kept the high information density. Showing it to others without explanation revealed that the amount of data was getting in the way of the core message.

Final version: back to basics

In the final version, I made two simplifications: the three groups have been reduced to two and the individual data have been added together for both groups. While losing details in this sense, this data presentation shows more clearly how the PageRank of the two groups is distributed, and makes instantly clear were the peak is in both groups. It allows the viewer to compare them side by side.