16 Décembre 2019
16 Décembre 2019
It’s in the nature of human beings to want to see things in advance: anticipating, forecasting, prophesising... In every area of our personal and business lives we’ve created all sorts of ways to reassure ourselves about what will happen next. From weather and economic forecasts to superstitions and religion, all of these rely upon some kind of understanding of the past and what’s happened before.
We increasingly keep track of this past using data. Data in all shapes and forms, not necessarily just dry numbers and figures in a crap Excel table. The past is our collective history, composed of a collection of testimonials, images and stories. Yet we often rely on these crappy numbers to say definitively, based on what we’ve learned, that something might happen. According to a survey done by KPMG in 2016, (and an article in Analytics-Magazine.org by Will Towler) 80% of organisations "report more accurate decision making when using data visualisation tools. Further, 86 percent of companies report faster decision making through data visualisation, while 81 percent noted that the tools improved customer behaviour insights." (Links: Survey and Analytics.com)
"Data visualization should facilitate decision-making, a goal that is difficult to achieve for many"
Based on these assumptions, I asked myself a few questions to see if and how data can be helpful.
Among all those questions resides a broader question about how we communicate with data, how it’s perceived and what it can tell us.
In our digital world, we use the term "big data" to talk about the information available from our computers, mobile or not. (In truth, this is just data that can be easily collected and analysed from many devices.) In its raw form, data does not mean anything. In fact, it’s pretty useless unless you put it in the right order, and more importantly find a meaning in that order.
A great example of how data can be used now is the TV series House of Cards. This, the first series released by Netflix, changed the way people create TV programmes. Before that, networks relied on instinct and pilots to guess what people wanted, and how they reacted to a new series. This was based on gut feelings, trends, and senior producers who had a good understanding of where the wind might blow.
This all changed with House of Cards. In this case, Netflix based its new series on evidence gathered by data analysis to create a targeted offer. This showed that people had a great interest in politics, that they like unconventional characters, and that most them liked Kevin Spacey. The story was then built around these insights into what the audience might like.
Moreover, to make sure their audience was targeted correctly, they created more than 10 different trailers to pinpoint the exact user needs. (Source: Gigaom) Even the way the episodes were later released took into account the fact people liked binge watching, so the entire series was released ahead of a long weekend, generating millions of conversations about the series–and ultimately creating lots of new subscribers to the Netflix channel.
Netflix made a great step with that move, using data to highlight how to build a narrative designed for its audience. They used this data not only to measure past audiences, but also to predict what to do next and how to write it.
We have millions of weather stations, collecting a tremendous amount of data: wind speed, temperature, humidity, atmospheric pressure etc. All that data, collected in real time, is sent to supercomputers that then transform it into the small clouds or nice yellow suns we see on our smartphones.
Even before the digital revolution, TV and radio presenters provided commentary on the never-ending story of weather: how sunny the weekend ahead might be, the likelihood you'd have snow for your next ski holiday, and risks of sun burn, if any. The beauty was in the commentary’s playful theatricality, combined with the lack of accuracy behind it.
Even if recent innovations in forecasting have improved accuracy, one thing we can predict for sure is that the weather always proves itself to be fundamentally unpredictable. And this is where the human touch is important. The best weather forecast can't predict how it will affect our feelings: how sun will brighten our mood; how the rain will dampen our spirit. Weather forecasting is an area where we cannot rely on data alone, where we need human storytelling through presenters or richer visual displays. We like these fun representations of the weather not because they’re more helpful, but because we enjoy making our own mind up from that information.
Data science is inescapably complex. In fact, put the word "science" in front of anything and you probably expect a tough subject. As a creative constantly driven by visual outputs, as soon as I hear "science", terrifying thoughts of nuclear plants, spacecraft, medicine, or computer motherboards spring to mind. But I’ve discovered there’s really nothing to be afraid of.
The bad publicity around data science is probably due to all the jargon. It may help the experts, but it doesn’t help the rest of us. And most of the time, with data, it only has a real value if it’s communicated well to an audience. We’ve all experienced that moment when, during a presentation at work, you reach the page full of Excel tables and charts and you realise you don’t have a single clue what the presenter’s talking about.
As a counterexample of that, I invite you to look at the diagram from "Information-is-a- beautiful-war.com (Edit: Now available here: https://youtu.be/2U1dsMGHogw)
This is a timeline of Earth's average temperature since the last ice age glaciation. Does it feed you with a long list of numbers?
Does it explain everything with jargon and scientific words? Nope. Instead they tell you some human history, and, as you scroll down the page, the imagery lets you draw your own conclusions. Have a look, and ask yourself this question: Is it convincing? Do you really think humans have nothing to do with increase of temperatures?
The beauty of this data visualisation reaches the three points we always want to reach: informing an audience about facts through data, provoking an emotional response, and compelling people to take action. Now who wouldn't want to not pay attention to climate change?
When data fails to deliver the right message it’s mostly because people are being presented with just the raw data, explaining what has been done, instead of communicating to the audience what it means and what they should know. By doing so data creates discomfort, doubt and probably frustration. Nobody wants to be in a situation like this. So, we need to tell a story: we need to have a clear narrative that reaches a point we want to make. This has to take into account who we want to convince, and what is expected from the audience. By doing so, we share knowledge in a way others can understand and we empower the audience; that’s all we need.
What we want is to put people in a situation where they will understand and feel confident, and moreover will know what to do next.
We want our audience to be able to make better decisions and, armed with proven figures, they will have the right tools to do this.
"The Greatest value of a picture is when it forces us to notice what we never expected to see"
Creating a strong visual narrative is crucial. This is where we turn data into value and meaning, with engaging representations used as a key tool for understanding.
This is where data could really make a difference. Data not only tells stories but also takes the narrative to another level. In that regard, there’s an interesting course that started recently at Washington University about bullshit. The course teaches students how to detect potentially false information, and how to understand why and how it has been communicated. One of their examples is about the last keynote speech Tim Cook gave, when he talked about the latest iPhone sales.
(More about this here : https://youtu.be/q94VJ3KToK8) (Goto 3 mins)
On his diagram you can see sales reaching a high level with constant progression, using numbers based on cumulative sales. But if you look at the real numbers, what you'll see is a different story, where iPhones sales are actually decreasing. So is the data wrong? Is it telling lies? It all depends on the narrative you build around it. But more importantly, it depends on your point of view.
One of the problems resides is that the human brain remembers easily images more easily than text. According to a survey done by 3M, people process images 60 times faster than any other kind of information (3M Survey quotes and FastCompany "Why we remember more images"), and we almost unconsciously combine this with other logical information to form something that makes sense.
(Source : FastCompany)
When data is represented visually it uses codes, colour representation, structured layouts, scale, associations or symbolism and composition. It creates not only a logical representation but also a harmonious representation. In fact it's only by using all the strength visual design has to offer that it can communicate information easily and quickly. In essence, data is visual, otherwise it's just useless numbers.
Look at the work of http://krisztinaszucs.com to see how far these experts go into the representation of data. Each of the diagrams already means something, they already tell stories, but there is an extra addition to it: the visually appealing representation of the subject.
No one can make a decision just by looking at a series of numbers and charts. They need to be explained, and a context needs to be added in order to make sense of it all. To this end, each data story should have a meaning and a point of view.
"If the statistics are boring you've got the wrong numbers"
Data calculation isn't the problem anymore; collecting and storing data is cheap and easy to access in the cloud. If we want to use data successfully, we have to bring it to life. Data is inherently visual; data exists if it is represented.
Data does not lie, but the narrative might….Publié par
Head of Experience and Creative Lead at FOVE By Sia Partners