Using {Big] data to reveal stuff about how we are



Screen Shot 2013-03-19 at 5.25.09 PM-2

Interesting piece covered in a lot of mainstream press today by Alberto Aceri of Bristol Uni and team about trends in words used in english Language novels over the last 100 years reveals about us:

" Britain's literature has grown less emotional since the 1960s, but American literature has become more so. Overall, English-language literature has used far fewer emotionally-charged words over time, but American writers have bucked the trend: They've ramped up their use of "mood words" in the past few decades as Brits have grown more stoic." With the exception of "fear" words (as the red line above indicates).

The ebb and flow of sentiment is also clear:

Screen Shot 2013-03-19 at 5.59.12 PM

The point being that even really simple analyses of data that's lying about (or observations that can be turned into data) can reveal important changes in our culture

NB THIS may or may not (as the authors and commentary in the Atlantic both observe) reflect how we experience things: e.g. "the socially-conservative
mores of Elizabethan England led to an increased demand for writing ''obsessed with romance and sex"" BUT it certainly reflects the culture and the products of that culture. And it's free!

So today's challenge is this: what can you do with the offcuts of data you have lying around to reveal what's going on beneath?