Data Has Feelings

Writing project. Thinking about how data makes people feel, and how the way people feel shapes data.

[data-has-feelings]

The absence of fear: Martin Orne's strange experiments

If you’re collecting data from people, how they feel at the time is likely to affect the data you get. This might seem obvious, but it gets overlooked all the time. And so in the 1950s, a psychologist called Martin Orne devoted a strange interlude in his career to making this point to his colleagues.

He was doing it mostly because he was deeply skeptical about popular claims at the time that hypnosis was a really powerful way of getting people to behave differently. Frankly, I think he also did a lot of it because he thought it would be funny, and also because ethnical review boards hadn’t been invented yet.

His basic premise was that people behave differently in psychology studies than they would in real life. One of his demonstrations of this involved sitting a group of students down for an experiment (he told them) in which they had to solve a page of maths puzzles, tear up the page and throw it in the bin, and then get another page of questions and start again. That was it: they just had to keep going until the experiment was done.

"Continue to work," Orne told them. "I will return eventually".

Several hours and no further input later, his students were still plugging away at their pointless task and yet were showing, as Orne put it, "relatively little sign of overt hostility".

It sounds ridiculous because it is: people were doing something meaningless for hours, just because somebody asked them to. Orne's contention was that while people wouldn't normally do something like that, being asked to by a psychologist, as part of an experiment, was an entirely different matter.

In other experiments, he asked people to do bizarre and dangerous things, like sticking their hands in jars of acid or handling venomous snakes. Some of his participants had undergone "hypnosis" beforehand, but his control group hadn't – and crucially, it didn't make much difference. Most people were happy to do as they were told.

Orne argued that people expect different things from psychology studies than they do from real life, which means evaluating behaviour-changing interventions like hypnosis in the lab isn’t useful. His experimental subjects trusted that because of the setting they were in, the acid would be fake and the snake harmless, and acted accordingly.

They were wrong to trust him, actually. Wholly unnecessarily and for reasons best known to himself, Orne used a real venomous snake with a glass screen that would come down if his participants reached for it. The acid was real, too – if participants touched it, they immediately had it washed off.

Put people in a research study and the special context means that they won’t always feel the things they might usually feel: fear, trepadition, or “overt hostility”. Those are all threat responses, justifiable when faced with snakes or infinite maths homework in the wild.

But people signing up for psychology studies of their own free will generally feel safe doing so, so there’s no need for that threat response. A different set of feelings might emerge instead – competitiveness, maybe, or curiosity. Different feelings lead to different behaviours, and different answers.

And potentially a snake bite, I guess, if the glass screen fails.

[data-has-feelings]

Together and alone: Treating desire paths as data visualizations

Accidental desire paths...

I get a bit obsessed with the idea of desire paths as natural data visualizations. Thinking about them in that way taps into lots of the things I’m interested in, and that makes it hard to know where to start.

Let’s start with what they are. ‘Desire path’ is the name given to that thing where there’s a trace left on the landscape from lots of people all behaving the same way. The classic example is a path worn diagonally across a lawn because it’s the quickest way across.

They’ve become a bit of a meme: Tumblr loves them, and so does Reddit, and there have been articles and podcasts all about them.

In the design world, desire paths are often treated as a metaphor for a way to approach design: look at what people are already doing, and then do something to formalise and solidify it. Architects have done this in the literal sense, by letting people wear their own paths across lawns and concreting over them afterwards.

But the term 'desire path’ also end up getting generalised to other, less useful things, like the ones in this post. Dips worn in stone steps by decades or centuries of feet. Paintwork eaten away by the acid touch of thousands of hands. Spending any time looking at traces like these makes it impossible to keep feeling unique or alone.

I had an experience like this in the first Covid lockdown. I had a path that I’d take on my regular walk up on an artificial hill at the edge of my local park, through some long grass and some trees and away from the slightly more crowded lawns below. I thought I was the only one doing it, but within a week or two a little path had appeared in front of me, and I didn’t know how to feel about it. Was I wearing a desire path all by myself? Or were there other people doing the same walk as me, and I was being less original than I thought?

Is it comforting to know other people are sharing your experience, or discouraging to know that even your morning walk isn’t uniquely yours?

I’ll tell you what else: it’s very hard to ignore a desire path, once you’ve seen one. Nobody cuts across the grass slightly to the left of an already-worn path just because they can; everyone’s feet just sort of fall in line with the easier route, where others have gone before. We’re communal animals, and we trust that the wisdom of crowds will get us there.

I don’t think this feeling is limited to these kinds of physical, incidental records of what everyone else is doing. In a world of infinite data collection and production, we’re constantly bombarded with images of what everyone else is doing. We’re aware of what 'average’ is and whether we’re above our below it; we know how we rank against everyone else in our fitness apps and our music apps; we’re constantly told exactly how many other people share our opinions and identities and voting habits.

And, genuinely, that isn’t always a bad thing: it can generate a sense of community and not-alone-ness, even with people we don’t know and can’t see. But it can also feel alienating, or push us towards conformity.

...and intentional ones

All of this is enhanced when the 'path' is a reflection of something intentional.

Like: there's a photo of me in Budapest, next to a bronze statue of a stout, mustachioed little man, my hand on his belly. The belly is worn and shiny from thousands of other tourists doing exactly the same thing, which means that I know for a fact that my photo is completely un-unique.

I didn't know anything about the statue before I stumbled across it (he's The Fat Policeman, installed in the eighties) or why people might rub his belly (there are various tourist-friendly narratives around the tradition). I just did it because it was evidently the thing to do.

[Source: Wei-Te Wong on Flickr]

There are places where similar practices feel more serious: gestures that are meaningful to the individual, but repeated to the point where they have a visible effect. Like a statue of St Peter in the Vatican that visitors traditionally touch the feet of, to the point where they're now smooth and worn.

Or Oscar Wilde's tomb in Paris, where devoted visitors have left a habit of kissing the stone with lipstick - to the point where there's now a barrier up to prevent them from doing it. That whole thing has been described as a gauche, annoying thing that tourists do, but I learned about it from this really good journal comic from ND Stevenson which makes me see it in a different way.

The quote at the end of the comic is, at least according to Tumblr, from Stephen Fry:

"Here's this man who believed when he died that his name would be toxic for generations to come. For hundreds of years his works wouldn't be read. He would stand for nothing but perversion. Utter disgust of a society that couldn't bear people like him... His tomb in Père Lachaise Cemetery, in Paris had to be restored because the polished stone of its surface had corroded through kisses. Wouldn't it just be allowed once to just wake him up for five minutes just to tell him that, then he can go back to sleep again?" 

But before we get too carried away by the beauty of our shared humanity, there's also the likes of the Molly Malone statue in Dublin, whose breasts are shiny from constant groping. That and many other similar examples just feel kind of gross.

So, it's possible to read all these examples as visualizations of What People Are Like, or at least What People Like To Do. But what am I supposed to take away from it all?

At minimum, it's interesting that people know they're doing the same thing as everyone else; the physical evidence is right in front of them. People find what they're doing meaningful precisely because it's a shared experience that connects them to other people who have gone before - even if that experience is repeating a stupid joke that wasn't all that funny the first time.

It's not all the same people every time, of course. And that's the thing to remember, I think. If you find yourself in one of these places, taking part in one of these little rituals, it's not just a moment that you're having with yourself - and you know it. You're declaring yourself part of a tribe of people, and signing up to their ideas about what the world is like, or what it should be like. So touch wisely.

[data-has-feelings]

Embarrassment: Why do I feel ashamed to be a Spotify outlier?

Embarrassment, when you break it down, is about feeling singled-out and different from everyone else but not in a good way. In data terms, nobody wants to be an outlier.

I think this is why I sometimes feel midly humiliated when I get my Spotify Wrapped at the end of the year. For the last couple of years in a row, it's told me that I'm in the top 0.1% of listeners for artists I like, and I'm still not sure how to feel about that.

As a society, we’ve decided we’re fine with the music we’re listening to being monitored and quantified by the company that provides it. We’re also fine with them packaging up what they know about us and sending it back like it’s a fun present.

But I’m old, is the thing. I grew up on CDs. And I don’t think I’ve ever let go of my teenage instinct to be really protective of the music I like, to want to keep it to myself, and oh my God, so what if I did listen to that song nine times in a row? It’s not like it means anything to me, shut up.

I also can’t help thinking about this ad campaign Spotify did years ago, where they pulled out data about individual users to make fun quips on billboard ads. Some felt weirdly personal: “Dear person in LA who listened to the ‘Forever Alone’ playlist for 4 hours on Valentine’s Day… you ok?”.

Which is exactly the nightmare my teenage self is freaking out about. Spotify knows that I'm Indie Artists Georg, and it's capable of telling everyone.

I don't think this is intentional, to be clear. Spotify clearly wants Wrapped to be a fun experience for its users, where we get to feel a bit special and a bit connected to the artists we like. But to me, there's a more sinister message in the background: "We know what you're doing, what you're like, and maybe even how you're feeling. Do you trust us to keep it to ourselves?"

*

PS: Maybe you're a well-adjusted person and you don't relate to this experience at all. If so, congrats. I suggest recreating something similar via this old-but-great widget from The Pudding which will connect to your Spotify and auto-judge your music taste for you:

[data-has-feelings]

The Mary Rose Museum (@maryrosemuseum.bsky.social)

On the anniversary of the sinking of the Titanic, we thought we'd answer a question that's often asked, "If they raised the Mary Rose, why

They're having dataviz fun, over on Bluesky, with a graphic that starts like this and - well, better to just go and have a look.

[brucebocchi]

colorblind pride flag just dropped

[data-has-feelings]

I'm going to call this a data visualization of a kind, because it tells you what this colourblind person experiences. It's amusing,* but the kind of amusement that comes with learning something you hadn't thought of before. For this person, colourblindness is not just about not being able to distinguish red and green, but also other colours that have elements of red and green in them. It generates empathy, makes you literally see through someone else's eyes,** so I'm saying it fits on this blog.

*To people who aren't colourblind.

*Unless you're also red/green colourblind, in which case this whole post probably just feels like an inside joke you don't get. Sorry. In case it's not obvious from the context: the individual lego bricks making up the different stripes in the Pride flag are not perfectly colour-matched.

[data-has-feelings]

Hans Rosling's hopeful bubbles

I've been writing about how data visualization can be an expression of hope, recently. And it'd be silly to keep on doing that without talking about Hans Rosling.

Hans Rosling is now one of those guys that it's embarrassing not to have heard of if you're a certain type of data person. He first became well known for his 2006 TED talk, The Best Stats You've Ever Seen, and quickly became synonymous with a certain type of evidence-based optimism. Here's the talk:

Rosling is a charismatic speaker, but the real star is his bubble chart - a style of animated chart that was everywhere for at least a decade afterwards.

In this video, every circle on the chart is a country, scaled to its population size and plotted by life expectancy.

Rosling is explaining that people have an outdated idea of what global health looks like. His chart shows how things stood in the mid-20th century, a picture that dominates people's perceptions even now.

But time is everything, and that's the extra dimension on his chart. He presses a button and time starts to tick forward, and suddently he's a racing announcer, keeping up a fast-paced commentary as the bubbles float upwards, life expectancy gets longer, and the world gets, broadly speaking, better. By the end, the audience is cheering.

Rosling spent the rest of his career making variations on the same point: that the world is often doing better than we assume it is, and we can find that out if we actually look at the data.

That talk was nearly twenty years ago, incredibly. Since then, it's often been very hard to think of the world as a place that's good and getting better.

Rosling was often accused of being too optimistic, and ignoring real problems in favour of his argument. His response was that he didn't see himself as an optimist, but a "possiblist" - a term of his own which he defined as someone willing to understand how things have improved so far, and take that as evidence that it's possible to go further.

The sentiment has been put more pithily by the founder of Our World In Data, a similar project which documents all kinds of global trends: "The world is awful. The world is much better. The world can be much better. All three statements are true at the same time."

It's a hopeful outlook, but a specific - and annoying - kind of hope. It's annoying because there's nothing in this data that shows that we can relax, that things automatically get better by themselves. It only shows that things can get better if we put in the work.

I've said before that people expect every chart to contain bad news, these days. But to expect bad news is to assume you already know the answers, and it's dangerously close to assuming that there's nothing you can do.

Our World In Data have a graphic on life expectancy, by the way, as they do for everything. And the answer to the question of whether it's improved further since Rosling addressed that crowd is - well, it would be nice if it was an uncomplicated yes, wouldn't it?

The real answer is: yes, dramatically in some places; less dramatically in others, where life expectancy is already high; but Covid deaths put a substantial dent in progress in some places.* Things don't automatically get better by themselves. Hope is, it turns out, complicated.

Life expectancy

The period life expectancy at birth, in a given year.

* By the way, a point that's not explained often enough is that "Covid-19 causing a fall in life expectancy" doesn't mean "the existence of Covid-19 means everyone will live shorter lives, forever". The measure of life expectancy that charts like these use are really just a way of expressing how many people died at different ages in a given year. That's why it can change over time, as circumstances (hopefully) improve and deaths (hopefully) happen later in life.

[data-has-feelings]

Hoping for daylight

It’s late January, one of the worst parts of winter. I always find this stretch the hardest: midwinter is technically behind us but there’s still a long, long way to go before spring.

At this time of year I start fixating on how much daylight we get in a day, because it’s the one thing that’s reliably getting better. It’s a habit I got into in the bleakest, dullest parts of the winter Covid lockdown. Nothing felt in my control, or even really in humanity’s control. So it was comforting to remember that neither I nor a virus can screw anything up so much that the Earth changes its trajectory around the sun. Lighter days are something to rely on.

This year, it feels like more people online are posting about exactly what point we’ve got to. We’ve reached the solstice, we’ve done the shortest six weeks of the year, the shortest ten weeks, keep going, keep going. One thing that’s reliably getting better.

There’s something about our fixation on the progress of the sun that feels prehistoric. Like the chamber at Newgrange, a neolithic monument in Ireland, that is aligned to illuminate perfectly on the winter solstice:

Well. I haven’t seen anyone build anything quite that good, lately. But people do still seem to like to make things that help us keep track and know where we’re going.

For starters, I highly recommend clicking around on good old timeanddate.com, where you can find a sun graph for your location, showing you exactly what trajectory you’re on to reach brighter days:

There are plenty of original charts going around, as well. On Bluesky, I’ve enjoyed browsing the work of Neil Kaye. Did you know, for example, that the days don’t all get brighter at the same rate? There’s a stagnant period in December and January where the length of the days don’t change that much – and man, doesn’t it feel like it? – then as we enter February, the change speeds up.

If you can understand this one you’re smarter than me, but still. I think it’s worth paying attention whenever you see one of these graphics, or some commentary on how much lighter the days are getting, or whatever. The message always seems to boil down to one thing: it won’t always be like this.

[data-has-feelings]

How to watch a baby

A digital comic by Kristen Radtke.

This comic from the Verge is a pretty incredible exploration of how much data new parents can choose to collect on their babies, and what it feels like and means to do that: whether it's about joy, control, proof of love, a barrier against fear, etc.

[imaginebetterfutures]

I made another sculpture! This one is kind of hard to document well, because I don't have access to a super clean white gallery space so apologies that the images aren't amazing. But I think it's a cool piece.

WICK is, in some ways, a clock. Slowly, invisibly, the water travels through the cotton fiber rope out of the glass vessel and down. Eventually, when it touches the paper, it will react with the small pile of powdered dye that's there, and create an image.

Here's what it looks like when that happens (sped up 12x):

I wanted to make something that engaged with ideas of motion, stillness and scale. At first glance, the piece is static. But what you cannot see with the naked eye is that there is quite a lot of motion happening — water being pulled, molecule by molecule, up and through the wick. It's simply happening at a scale that we can't easily see or engage with.

In a way, Wick is a time piece. Each image created is a portrait of a day — many hours of seeming stillness suddenly made visible on the page. The water travels through the wick at a set rate (about .1 inch per minute) and so you could conceivably use it as a stopwatch, or a timer. And yet, doing so seems absurd, because we want things to be precise, visible and immediate. Wick is none of those things.

In our class critique, one of my fellow students called the piece "frustrating," which I actually really appreciate, because to me that is one way to describe trying to wrap your mind around time and scale and motion — it forces us to be a little bit uncomfortable. To wait, and to look closely, and to wonder whether or not the scale and speed we operate at is the only one worth trying.

[data-has-feelings]

I'm reblogging this here because I always think of this kind of self-making art as a weird cousin to data visualization.

There are lots of other artworks that this reminds me of: Olafur Eliason has melted glacier ice on circles of ink, and Charles' Ross Solar Burn works track the sun's progress across the sky by scorching a piece of wood with focused light.

Like data visualization, these works are visual records that capture something that's happened in the real world. The 'something' that happens - in this case, the progress of water through rope and dye and paper, and the minute individual variations in those things that create a distinct image - is more complex and multidimensional than the representation you're left with.

But works like this are arguably far more faithful to the thing they're recording than most data visualizations are. There's no designer in the middle, picking and choosing and simplifying. The end product here tells you no more and no less than exactly what happened when this particular dye on this particular page got wet on this particular day. It's up to you how much effort you put into reading it.

Anyway, "frustrating" is a good word for this kind of work! Living in a world with so much information available can make us feel like we're entitled to know everything about everything, instantly. And we're used to treating records and representations as if they give us a comprehensive account of reality.

Slow visualizations can challenge both those things. A pattern emerges on its own schedule, without explaining itself, and we're made to admit we're not as much in control as we thought we were.

[data-has-feelings]

So I’ve been writing about people who mess with their survey responses for fun, and I think one of the reasons that sort of thing feels fun to do is because people think of statistics and data as very, very serious.

That’s not surprising. If you see a statistic in a news article or a post, it’s almost always telling you about bad news.

Covid made it worse (Covid made everything worse). We started bracing ourselves for every new data point, almost every day. And that flinching instinct is always there, whenever it’s time to hear more about climate change or inflation or any of the other quantifiably awful things happening in the world.

People who have stories to sell know this, and they work with it. If you want someone to take a problem seriously, you make sure you present them with a serious-looking chart.

And that means that serious-looking charts have their own visual language. We all know what they look like, and when we see them we switch off out of self-preservation.

But! Data visualization doesn’t have to be like that! There are innovative, joyful approaches to presenting data.

Maybe you’ve seen the work of Neal Agarwal, a guy who makes things online that are just fun.

Here are some things you can do on Neal’s website:

Spend Bill Gates’ money. Your choice whether you spend it on Big Macs, Apache Helicopters, or the Mona Lisa, but you’ll come out of it with a stronger sense of what a billion dollars is.

Take an elevator into space. Learn along the way about what goes on at various altitudes, and enjoy delightful touches like elevator music and a choice of scarf to keep warm.

Get the stats on exactly what’s happened in your lifetime, from inflation to the number of rotations of the moon around the Earth.

This is sort playful creativity is vital for coaxing people back into reality.

Because there’s nothing inherently miserable about finding out about the world through data. I’m not saying that it’s all good news. But that little burst of joy when you start playing with something like ‘space elevator’ is a starting point, a tiny pitch for the merits of paying attention.

Maybe after that point, after your attention was engaged, you started to feel other things in response to Neal’s games: wonder, or anger, or hope or sadness. But the point is, it was much easier to feel those things because your defences weren’t up already.

[data-has-feelings]

Sometimes, people mess with the census as a protest because the government is denying them human rights, sure. But an unavoidable problem in survey design is that people will also sometimes tell you complete untruths just because they think it’s funny.

Census Jedis are your classic example. In 2001, censuses in the UK, Australia and New Zealand asked about religion for the first time. 2001 was also the early days of widespread internet, and so a meme began to circulate in chain emails.

People were urged to put ‘Jedi’ as their religion on the census, partly for fun and also (they were told, though this isn’t actually true) because it would force their government to recognise it as an official religion.

People did. In England and Wales, 390,000 people did, outnumbering several minority religions.

And why? Some people framed it as a protest against asking about religion on the census, or against the census in general. And sure, there’s an anti-establishment sort of feeling behind it.

But for my money, people mostly did it because it amused them. Because surveys are boring, and people get very serious about the census in particular. Phrases like ‘civic duty’ get thrown around. And there’s a form with all sorts of serious boxes to tick, but also a free space to write whatever you want, and, well, it would be funny, wouldn’t it? It just would.

There are consequences to these sorts of shenanigans. Questions were asked about the quality of census data. Atheist and Humanist organisations begged non-religious people to stop doing it, so as not to artificially inflate the count of religious people. Sociologists have written papers.

But all of that just makes it funnier, doesn’t it?

Look, I’m not encouraging doing this. Nobody who works with this sort of data in real life can afford to – we’re all terrified that the rest of you will decide something else is funny next, and muck up a carefully-calibrated data series we’ve been using for years.

I’m just reminding everyone to beware. There’s no power on Earth that will keep people from being silly, including within your datasets. Sometimes you just have to work with it.

[data-has-feelings]

This is a return from the 1911 census in England and Wales. At the time it took place, the movement for women’s suffrage was well underway, and this was the result of an organised campaign to use the census as a vehicle for protest.

In many ways it was ideal as a form of passive resistance. A census is supposed to collect data on everybody. The government was spending huge amounts of money and effort to get the details and whereabouts of every single household on one night in April. Why not use the opportunity to voice your anger about the status quo, instead?

So that’s what the Women’s Freedom League did. They encouraged women all over the country to evade the census entirely, or return their forms with slogans like “No vote, no census”.

Some had a lot of fun doing it: there was an all-night party at an ice rink in London, with attendees circulating between there and a nearby restaurant “to baffle the enumerators”.

Others had to put up with enraging condescention from within their own household. Like Eleonora Maund, married to a no doubt delightful man 31 years her senior. He wrote down her details; she crossed them out. He wrote her details down again in red ink, and lodged a protest against her protest: “My wife … attempted by a silly subterfuge to defeat the object of the Census. To which as ‘Head’ of the family, I object”.

I made this blog to talk about how people's feelings will get into your data collection, whether you want them to or not. This is a prime example. The British government in 1911 felt it had the right to ask all its citizens for information about themselves, for the worthy project of compiling statistics that would help it make good policy. But the suffragettes felt differently. They were angry, and they didn’t see why they should co-operate with the work of a state that refused to give them a democratic voice.

So they didn’t. Because at the end of the day, a census form is a blank piece of paper. And if you give someone a piece of paper, what’s to stop them from telling you what they want you to hear?

[data-has-feelings]

Knitting for the quietly enraged

Yarn-based data visualisation – which is very much a thing, if you didn’t already know – often seems to come with an undercurrent of rage attached to it.

One of the first examples I remember was a woman in Germany, who knitted a colour-coded scarf that recorded every time train delays ruined her commute.

Not too long after came a scarf from a city councillor knitting her way through meetings in Montreal. She changed colours based on the gender of the person speaking: green for women and red for men. The scarf was, inevitably, mostly red.

These both feel allied with the slightly bigger phenomenon of the ‘temperature scarf’. Lots of people have done these, and they’re the ones you’re most likely to have seen. Artist and author Josie George knitted one in 2020, recording the weather every single day for a year.

Of course, the weather rarely behaves as it should these days. After recording a week of unseasonably hot May weather, she tweeted:

“For me, a month of deep grief and deeper resolve, side by side, as I looked unflinchingly at the world's damage. I read and spoke words of hope and change as I knit. I continued. I committed. I began again.”

Knitting is steady, continuous, and pattern-driven. It’s perfect for making a record of things as they happen.

And that’s how it goes with a lot of things we get angry about. They happen once - a train is late, a man drones on, the temperature lurches into something alarming - but it’s just one little thing. Not enough to justify rage.

And then they keep happening, and happening again. React to any one instance and you look irrational. But keep a record, quietly, politely, and calmly, and maybe you get to feel like you’re building the evidence base for a revolution.

Or not. Whether or not you get to make your point, whether or not other people listen, just the act of recording might feel important. You get to process the things that you have to put up with, without ignoring the part of yourself that’s insisting it isn’t right.