I just finished watching Hans Rosling’s documentary for the BBC “The Joy of Stats”.
Hans Rosling is the ebullient swedish professor who came to widespread public recognition through his passionate TED talks where he used statistics to unravel and debunk many widespread myths and misunderstandings about world health and the third world.
The documentary “The Joy of Stats” is a one hour whistle-stop tour of the history and power of statistics to predict and understand our world.
And it got me thinking.
Rosling, like many others, predicts that visualisation has the power to transform the way we think of data and allow us to understand and interpret vast amounts of information.
“Visualisations tell us stories”, he says
Or do they?
Hans Rosling is a remarkable story teller. The TED talks which brought him to mainstream attention are a fabulous example of a man telling a story with three acts, emotional arcs and a wisdom all illustrated with pictures and animations. He finishes, leaving you entertained, excited and above all informed.
Here’s the crux. It wasn’t the picture, the visualisation, telling a story. It was Hans Rosling! I’m sure that if I had seen the data and the visualisations without his explanation, I would have only had the smallest amount of insight. You need lots of other contextual information to make sense of it all.
Now, I don’t want to say that this all a big waste of time. It definitely isn’t.
But as an individual person looking at some information that’s been visualised, what am I supposed to do with it?
I think real insight often comes with a razor sharp focus. In this case, Hans Rosling is the person with that focus.
This brings me to the second part of what struck me after watching the programme. It’s to do with the visualising data about a large group of people versus visualising data for one person.
Statistics developed out of a need to understand what society as a whole was doing. How many people are dying? What’s the average life expectancy? Is our population growing or shrinking?
The answers to those questions are used by politicians and policymakers to affect change on the societal level.
We’re using the outcome of statistics to pull levers to affect change to all of society to get an outcome that we want.
Something of the language that Hans Rosling used puzzled me. I think it’s been bothering me a while because lots of people do this.
It’s implied that visualising data is a way for ordinary people to understand and interpret data themselves, effectively giving them access to the same analytical machinery that once was only available to the few, in the process democratising the machinery of decision making.
I think this is an important and truly revolutionary goal but the way that we’re going about it is missing the mark.
Because we’re still tending to think about visualisations from the point of view of someone very far off the ground, seeing the actions of the people as little ants scurrying around.
To make visualisation a democratic tool we need to make visualisations that allow us to make decisions about our own lives.
And I don’t mean, “by looking at this data I can see that there is a strong correlation between smoking and lung cancer. I don’t want to get lung cancer. Therefore, I shouldn’t smoke.”
Correlations and statistics by their very nature can not say “if you smoke, you will die twenty years earlier”. It’s only a statistical link. So their ability to persuade anybody to do something is very limited.
So then how do we make a visualisation personal?
Focus on the concrete.
Let me give you a simple example. It’s not strictly a visualisation, but it illustrates the point quite well.
Let’s say that every house on your street is roughly the same size with about the same number of people living in them and imagine that each house has an electricity meter in it which records power usage and stores that information. Let’s say that there is data for the last year.
What could you do with that?
You could draw some graphs that would show the average energy use of a house on the street and show how it changes over the year. You might discover that there is a large spike during the summer. You look deeper and find out that’s because everyone is running their air-conditioner. That’s an example of using that data in a very top-down way, looking at the little people from above.
What if instead you could see whether your energy use over the last hour was lower or higher than the average of your neighbours’ ?
You could actually use that information in a very tangible practical way. You might discover that when your refrigator makes a noise and kicks in there’s a spike in your energy use and suddenly you’re using more energy than your neighbours.
Or let’s say you don’t make it about competition with your neighbours but rather with yourself. What if you just see if your energy use is lower or higher than your own energy use exactly a year ago? If you wanted to reduce your energy use you would have a really tangible and simple tool to help you do that.
I think this whole discussion also has an influence on the kinds of information we should be working towards getting governments to release.
Currently, the focus appears to be on the kind of information that comes from the top down approach. For instance, here in Australia, the gold standard for open government data right now is the Australian Bureau of Statistics that have for some time now released large amounts of census data in good machine readable formats available under licenses that allow liberal reuse. It’s amazing what you can find there.
But, it’s very hard to see how to create a personal visualisation out of any of it.
So, we have a challenge.
What data should we get from government which we can use to create visualisations that can allow a person to make decisions about their life?