This post is a review of Speak Data: Artists, Scientists, Thinkers, and Dreamers on How We Live Our Lives in Numbers by Giorgia Lupi and Phillip Cox. Well, it’s partly a review. It’s also partly an excuse to talk about the workshop I’m running with Sophie Sparkes next week, but I’ll get to that (I promise it’s relevant!).
I’m excited to review this book: I read it the day it came out a few months ago (those who know me will know how much of a hero Giorgia has been, in particular with the groundbreaking Dear Data project and her definitive words around the concept of data humanism. But it’s been far too long in the review. So I wanted to look at it in terms of one of the key threads she asks throughout. What is data? After over ten years writing a data themed blog where every post is a question, it seems criminal that we haven’t asked this question yet.
The premise is a simple one. Lupi and Cox interviewed 17 people – many of whom probably wouldn’t describe themselves as data people – and asked each of them the same question: what is data? The list includes a cardiologist, a composer, a museum curator, a meteorologist, an AI artist and the author of Atomic Habits. Quite a combination! Each conversation goes much further than that first question, but it was the definitions that really got me thinking. The book itself groups each of the 17 experts into five categories: Data and Identity (Counting Ourselves), Data and Communication (Making Meaning), Data and Health (Vital Signs) , Data and Technology (Machine Thinking), Data and Creativity (Paint by Numbers). Every category had its own preface with the thoughts of Lupi or Cox, along with data illustrations from Giorgia.
As is my wont, I couldn’t just read the definitions – I wanted to find similarities, differences and influences. The answers were sufficiently varied that I wanted to visualise them simply in some way. And once I’d lined them all up, I realised they weren’t really 17 separate answers at all. They sit on a spectrum, from data as cold, hard fact at one end to data as something deeply human at the other. Here’s how I’ve visualised it:
At one end is Eric Topol, the cardiologist, who gives perhaps the most stripped-back answer possible: “Data is data. I don’t know if I can get more atomic-level than that.” Seth Godin isn’t far away, describing data as “a mass of verifiable numbers that have no information in them … yet”. Hold on to that “yet” – I’ll come back to it.
Next come pragmatic definitions that I suspect some of us might have come close to, or might have picked up along the way. Adam Grant talks about information gathered through systematic and rigorous observation, Max Roser (of Our World in Data fame) about structured information about something that matters, and meteorologist Bernadette Woods Placky about “the grounding of fact with truth”. David Putrino adds that it’s any piece of information you can use to form an opinion or guide an act. Rigour, evidence, truth: all very sensible, and all very much at the “best practice” end of the spectrum.
But then things get more interesting (for me, at least). Composer Nico Muhly describes data as “information in context” – “points that invite a line to be imposed on them”. His analogy is a lovely one: scatter some marbles, ask people to organise them, and you’ll get lots of different systems. The data exists before the system, and where it gets artistic is in how you draw the connections. As someone who has spent a fair bit of time drawing connections in unconventional ways, this is right up my street.
Artist Sougwen Chung goes further still, thinking of data as “a material to work with, similar to paint or clay”. For MoMA curator Paola Antonelli it’s “a well that both art and design can draw from”, and for artist Ekene Ijeoma it’s simply “another way of seeing”. Andy Marra calls data a common language that a diverse group of stakeholders can understand. And Naresh Ramchandani gives what might be my favourite answer of all: data points are plot points, “irrefutable points of drama in a real story”. Something gets worse, something gets a LOT worse, something gets better. I’ll certainly be borrowing that one.
At the far end, data becomes personal. Bon Ku reminds us that patient data is “a representation of patients – but not a complete representation”, and Kate Crawford describes it as “a capture of the commons” (which rather begs the question of who’s doing the capturing). James Clear calls it recorded feedback – the feedback that would otherwise vanish – and “kind of a magical thing”. Refik Anadol thinks of data as a form of memory, one that only becomes human again when you give it back its context. And John Maeda simply says: “Life. It’s my lifeblood.” Hard to argue with that!
So the beauty of the book is that it allows anyone to be seen – think what your personal feelings are about working with data and how you define data as part of the process, however you might work with it, and you will probably find something that aligns with your own views, inspiring you to work with data in the same way that many of these visionaries and experts do. And the commentaries in between show us how important all five categorisations of working with data really are, and how relevant to all of our lives, not just our jobs.
Additionally, I love the interview format. In this day and age it gives you confidence that you’re reading the verbatim words of real people, not just AI!
So I thoroughly recommend it to creatives in data, but on the back cover, it’s described as a thoroughly interdisciplinary study. That hits the nail on the head, there’s a lot bigger audience here than just data visualisation practitioners and data artists.
But the part of the book I keep coming back to is right at the very end. Over the last two pages, Giorgia Lupi sums everything up in a short guide to how to speak data – nine principles, each with a paragraph of explanation. And I couldn’t help noticing that the list starts with questions and ends with questions. You’re reading this in a blog called Questions in Dataviz, you can read all about Data Humanism in a book called Questions in Dataviz, and I am about to deliver a creative workshop called From Questions to Creation. So I liked this, a lot!
Here they are, summarised (and anglicised) in my own words:
- Start with questions – good data starts with good questions. Think like a detective: what are you curious about, and where’s the human dimension in the numbers?
- See the forest, count the trees – all big data starts small. Aggregates give us the full picture, but it’s the individual data points that capture what’s really going on. Both matter.
- Look for context – a data point is like a photograph. It might be beautifully composed, but what’s lurking just outside the frame? Modern data collection often edits out nuance and irregularity, and we need to welcome it back in.
- Embrace the unknown – what we don’t know can be just as important as what we do. Missing data is still data.
- Value accessibility, not simplicity – resist the temptation to equate simplicity with perfection. Life is messy, and the best data gives us reality “in all its untidy abundance”.
- Find the story in the numbers, and the numbers in the story – stories give numbers emotional resonance, and numbers anchor stories in truth. Treat qualitative and quantitative information with equal enthusiasm.
- Stay in control – more data isn’t necessarily better data. Decide what matters first, then use data to capture it.
- Use your fullest vocabulary – when visualising, start with meaning rather than numbers, and don’t let graphic conventions deaden your ingenuity. Find the balance between “pragmatism and poetry”.
- End with questions – numbers are human made, and “data-driven” doesn’t mean “unmistakably true”. Data literacy isn’t enough; what we need is full data fluency (a phrase very close to my heart, given what I do). Let data be “the beginning of the conversation”.
I’m not sure which principle resonates most with me – perhaps it’s principle 2: See the forest, count the trees. This really encapsulates my love of disaggregated data and focusing on every data point. But it’s the full list that shows to me the importance of questions when it comes to creativity, much more than answers.
Which brings me back to Seth Godin’s “yet”. That gap between a pile of numbers and something that actually means something is where I think a lot of people get stuck. Most of us are pretty comfortable at the “fact” end of the spectrum. We’re far less comfortable picking up data like clay, or drawing our own line through Muhly’s marbles.
And that gap is exactly what Sophie Sparkes and I are tackling in our workshop – From Questions to Creation – on Tuesday 3 November in London. I’ll be opening with a creative inspiration session, and I’ve been busy sourcing datasets for everyone to play with. Then Sophie will lead a rapid idea-generation session, taking everyone from a question to a creation of their own. Much more info is on the sign-up link, we’d love to see you there, and I’m happy to answer any questions you might have.
So, to return to the title question (something I admittedly don’t always do!): what is data, really? Having read 17 answers, I think the honest answer is “it depends”. Not just because it’s my favourite answer to most questions. But it’s better still to say that the answer is your own definition, and that definition could fit anywhere on a spectrum. Personally, I want to feel closest to “Paint by Numbers” and most of those lie in the middle section, that data is material. It’s the medium for creativity, for art, and for more questions. But the human sides of the definitions resonate strongly too. The answer doesn’t matter in the grand scheme of things, but taking the time to think about the question opens up a whole realm of ideas.