Friday, 28 November 2008

Data Visualization

I have been exposed to and been responsible for some ugly charts in my time and I’m really trying to get my head around data visualization at the moment….i firmly believe it’s how we win in the future, particularly with digital and all the data the cloud will throw up.

Right now, I am loving Edward Tufte’s work in this area. This guy really is wild; his seminal piece is right up there with Nirvana’s “Smells like teen spirit” or Public Enemy’s “It takes a nation of millions to hold us back” – In 1983 he released "The Visual Display of Quantitative Information"…ok it was a book but it was just as controversial. Tufte described himself as being dedicated in turn to "pictures of numbers," "pictures of nouns" and "pictures of verbs". Needless to say I highly recommend his work, and the drugs he was taking.

Tufte tells us that this is “the best statistical graphic ever drawn”!!

I think the chart works at an emotional level. It portrays the army of 600,000 on the left hand side as a powerful river that dwindles to a stream as Moscow is reached on the right hand side. Then, the stream reverses course when Napoleon retreats, and finally becomes a trickle in the end.

Napoleon's Grand Armee was ruined by a combination of extreme weather, disease, starvation, the Cossacks (who basically smacked them down) and the Russian tactic of burning villages before Napoleon arrived, denying the troops food and shelter. The final casualty rate was 97.7%.

In the 25 years since Truft wrote VDQI technology has advanced us in ways we never expected. Truft updated his thinking in “Visual Explanations” in 1997. Even the cool digi kids are banging on about the idea of networks, I constantly hear about “Influencer Network Analysis Charts” - which is “incredible, fantastic…particularly for word of mouth”. Typically these charts look like this:

This, believe it or not is the hierarchical structure of the internet – but what does it tell us? More importantly, what can we do with it?

Recently I stumbled across some really cool stuff. This comes from informationarchitects in Tokyo. It’s called Web Trend Map 3

The map pins down nearly 300 of the most successful and influential websites to the greater Tokyo area train map. This works for me as I am already wired to look at maps like these, I expect connections and popular interchanges (Think Times Square and Kings Cross St Pancras) – I get the information quickly as I don’t have to decode the chart. How many times has a client asked you “So what’s on the x-axis? ”

So imagine a world in which all the work that we did, all those complex concepts we generate and turn into business driving solutions for clients were delivered in much more user friendly formats? We could use everyday items and morph them to give our clients ideas that stick. I can see clocks, computer keyboards, weather maps, restaurant menus and credit cards as presentation devices in the future

It’s long been my professional mantra to take this thing we call ‘marketing analytics’ and make it more accessible. Don’t get me wrong, I love to talk through all the statistical tests over drinks with the rest of you; it’s just that our clients don’t. For the $1tn spent on marketing we operate in a business that barely nets us all $1bn (yes, that’s the whole business…including the data providers and the management consultants) – of course you’ve already worked out that only 0.1% of marketing spend is proven to be accountable. Surely we have to shoulder some of the blame for that?

In Tufte's world, "clear and precise seeing becomes as one with clear and precise thinking." We need clarity and precision more than ever. To think intelligently about anything worth talking about today requires an acute understanding of statistical evidence - whether it's the trade-off between budget deficits and taxation, global warming or spiralling medical costs. So yes, we are the most interesting people at cocktail parties

Understanding our world demands understanding numbers - Under these circumstances, Tufte and informationarchitects’ work can act as a machete to chop through the data jungle.

Friday, 21 November 2008

The End of Binary Thinking

Barack Obama’s win and all the discussion about a “Post-Racial” era in America made me think that we might be entering the end of binary thinking. Which is the why i am titling my new blog "the end of binary thinking"

The fascinating element about Barack Obama is that he is not easily classified in simple, binary terms, especially in regards to his race, upbringing and personal history. He is less the one or the other but sort of like me, a mix of a lot of different things. He is black and white, he is American but he grew up in foreign countries, he is a guy from Chicago but he only moved to Chicago in his twenties, he is an elitist and intellectual but he was a community organizer as well
Most the analytics teams I work with or those guys i meet informally as frenemies think and model in binary terms. It’s either 1 or 0, you are either in segment 1 or in segment 2, the campaign has either worked or not. The origin of this approach resides in the long tradition of philosophy of dualism and binaries. It’s tough to snap out of that, especially for mathematically trained brains who believe that everything can be reduced to a 1 and 0 formula, as any computer program is proof of.

But life is rarely truly binary, and the models that we are building should attempt to reflect this - life more accurate by moving beyond a simple binary structure. It’s often more useful to think less in opposites than in a continuum of elements, less in strong colors than in shades of something. Things are rarely true of false but more often a bit truer or a bit more false.