Holger Kuehnle: As a design­er work­ing with tech­nol­o­gy, I’ve always been real­ly pas­sion­ate about tech­nol­o­gy. I always saw tech­nol­o­gy as some­thing that empow­ers peo­ple, that helps peo­ple com­mu­ni­cate bet­ter, that helps peo­ple under­stand infor­ma­tion, and with that build knowl­edge and remove bias and bring us all togeth­er as a society. 

Aerial photo of the low turnout at Donald Trump's presidential inauguration

And then this hap­pened. Someone who seem­ing­ly no one wants to see become President became pres­i­dent. People in the UK—the United Kingdom—became very very afraid about the sheer num­ber of peo­ple liv­ing in Turkey and what would hap­pen if Turkey would join the EU, [what] that would mean to their social secu­ri­ty sys­tem. 76 mil­lion peo­ple in Turkey. And it prompt­ed them to vote to leave the EU. People start­ed get­ting very very very angry and agi­tat­ed about issues such as how much the United States spends on for­eign aid. And so, I’m look­ing at this and I’m not even quite sure what this means. Three bil­lion less, twelve bil­lion more. To me as and indi­vid­ual, and I’m assum­ing this goes for a lot of oth­er peo­ple on the stage here—for most of us—any bil­lion amount of amount of dol­lars sure sounds like a lot of dol­lars to me. 

So what’s going on here? In the first exam­ple, we’re look­ing at how many peo­ple attend­ed an event dur­ing a cer­tain spe­cif­ic moment in time, and we’re using that—we’re spend­ing a lot of media time—talking about how many peo­ple were there, how many peo­ple weren’t there, to make an assump­tion about whether this per­son is a legit­i­mate pres­i­dent or not. If you think that this per­son should not be pres­i­dent, your moment to do that as an active cit­i­zen was like­ly dur­ing the elec­tion. At this point this per­son­’s already stand­ing there up on stage get­ting inaugurated. 

And in the sec­ond exam­ple, we’re using a very large num­ber, 76 mil­lion peo­ple, cre­at­ing assump­tion that Turkey join­ing the EU would mean that all these peo­ple could sud­den­ly come to the United Kingdom and have a neg­a­tive impact on the social sys­tem there. 

And in this third exam­ple, we’re talk­ing about increas­es or decreas­es to a num­ber. A num­ber that’s actu­al­ly not men­tioned here. I don’t actu­al­ly know from look­ing at these head­lines how much mon­ey is being spent on for­eign aid over­all. And how has that been the past cou­ple of years—and prob­a­bly most impor­tant­ly so, for me to have a real­ly good opin­ion about whether that’s good or bad, what are we using this mon­ey for? What is actu­al­ly for­eign aid used for? 

So what do these exam­ples have in com­mon? They take a sin­gle num­ber out of con­text and bias toward a spe­cif­ic con­clu­sion that actu­al­ly fuels divi­sive­ness. The key here is con­text. You need con­text. And I’ll talk a lit­tle bit about how con­text is nec­es­sary to real­ly under­stand data, nec­es­sary to under­stand num­bers. And I want to do that with an example. 

So let’s imag­ine some­body shouts, Ball!” What does that mean, ball? Should I duck, should I be scared? Like am I going to be hit by a ball? Am I on a sports field, should I hold our my hands and catch a ball and that’s a good thing? I don’t know. I have no idea of know­ing what ball means if you just say ball.”

Is it a ball that I’m sup­posed to kick? Is it a ball that I’m sup­posed to uh… Either way, it’s sup­posed to be a foot­ball as well. 

Is it a small ball? Is it a large ball, a big heavy ball? I need a way to quan­ti­fy what this ball is. What does this ball mean to me from a quan­ti­ta­tive perspective? 

Line drawing of a tennis ball near a person's head

Well, how large is it com­pared to you how large I am? If it’s sort of you know, slight­ly small­er than my head, okay. I under­stand that. That gives me quan­ti­ta­tive con­text. That’s one of the key ele­ments of con­text I need to be able to quan­ti­fy what it means. 

Line drawing of a tennis ball near a person's head, showing the arc of the tennis ball's travel

And then next what’s essen­tial is under­stand­ing the his­tor­i­cal con­text. Where has this ball been before, where is it right now, and with those two com­po­nents I might infer with some amount of con­fi­dence, or less, where it’s going to be in the future. Is it fly­ing towards me? I need to under­stand the his­tor­i­cal context. 

And then next it mat­ters a great deal about where all this is hap­pen­ing. This bal­l’s fly­ing towards me. Alright, I under­stand. It’s sort of a small ball, slight­ly small­er than my head. I under­stand a lit­tle bit more what ball” now sud­den­ly means. 

Line drawing of the person, smiling, hitting a tennis ball with a racket

Well, if I’m hold­ing a ten­nis rack­et, and I’m on a ten­nis court, that’s a great thing. Ball’s fly­ing towards, fly­ing towards my ten­nis rack­et, every­thing’s great. 

Line drawing of a person, frowning, as a tennis ball breaks a nearby window

But if the envi­ron­ment that the ball exists in changes. If let’s say I’m stand­ing on the oppo­site side of a win­dow and that bal­l’s fly­ing towards me, that’s kind of a bad thing. Like I’m not going to be very very hap­py about that. 

So, the rela­tion­ships that this ball exists in mat­ters. It’s the rela­tion­al con­text that the ball exists in. So in order to under­stand con­text we need to quan­ti­fy, we need to under­stand the his­toric con­text, and we need to under­stand all the rela­tion­ships that what we’re talk­ing about exist in.

So how does all that apply to num­bers, right? I’m stand­ing up here, sup­posed to talk about data. Well, I’m gonna give you a cou­ple of exam­ples of a project that I’ve been work­ing on as a design lead at Artefact with one of our clients. And that project is called USA Facts. USA Facts is an unpar­ti­san plat­form for aggre­gat­ing gov­ern­ment data, with the goal of mak­ing that data eas­i­er to under­stand, more acces­si­ble, with the over­ar­ch­ing mis­sion to dri­ve and fos­ter more civ­i­lized pub­lic dis­course that is based in facts. And what USA Facts does is it takes all this gov­ern­ment data, puts it into a for­mat that is uni­ver­sal to some extent and can be ana­lyzed and then pub­lish­es it on their web platform. 

It does that with­out any form of inter­pre­ta­tion. So USA Facts does not take sides. And that’s actu­al­ly quite chal­leng­ing. And if you have to imag­ine like, this data’s actually—you can today go and access all sorts of gov­ern­ment data. It’s avail­able. However it’s all caught up in dif­fer­ent reports. It’s being pub­lished by dif­fer­ent agen­cies that have dif­fer­ent sched­ules of when they pub­lish these reports. It might be in a spread­sheet some­where. And so USA Facts does all this work of dig­ging this data up and mak­ing it comprehensible. 

We’ve work­ing with them for a while and we’ve built visu­al­i­sa­tions in which we’re try­ing as much as pos­si­ble to put data in con­text. I’m gonna show you a cou­ple of exam­ples of how we’re doing this. 

So, maybe we’ll go back to the pre­vi­ous exam­ple when we talked about for­eign aid. Remember we talked about three bil­lion less, twelve bil­lion more—well, we don’t know how much we’re spend­ing in total in for­eign aid. These are 2015 num­bers right now. This is a visu­al­iza­tion which breaks down the spend­ing, the total spend­ing of the United states across dif­fer­ent areas such as defense, wel­fare, and sub-units of those, of those spend­ing areas. So here we can see that in 2015 the United States spend $5.7 tril­lion. That is actu­al­ly real­ly a lot of dol­lars now, right? 

And so when we go over here and we scroll over a lit­tle bit to the right, we can see how that spend­ing is being bro­ken down. I can see it in rel­a­tive con­text to child safe­ty, defense…and, down here, for­eign aid. And we can see that in 2015, the United States spent $48.5 bil­lion. That is real­ly a lot of mon­ey as well. But if you look at it in rela­tion to the total spend­ing of the United States, it is only .86—making sure I have my num­bers right here—.86%. And that allows me to much bet­ter quan­ti­fy whether I think that the three bil­lion of twelve bil­lion more is bad, in con­text. Because I’m able to quan­ti­fy it, to have con­text that I can now see. 

Let’s take a look at anoth­er exam­ple of estab­lish­ing con­text in a visu­al­iza­tion. This one I’ll have to explain a lit­tle bit because it’s built for actu­al­ly being looked at in a brows­er so we’ll adapt it for stage a lit­tle bit. It’s actu­al­ly turned to the side because it’s meant for scrolling through. This visu­al­iza­tion shows the num­ber of peo­ple liv­ing in the United States that were not born in the United States—the foreign-born pop­u­la­tion. And it’s actu­al­ly turned to its side because it’s much more immer­sive to scroll that way. And it allows you to see that num­ber over time, in jux­ta­po­si­tion to his­tor­i­cal events that might’ve hap­pened that are impor­tant to consider. 

Another impor­tant piece here is that we’re show­ing this as a per­cent­age, not as an absolute num­ber. That’s anoth­er way of quan­ti­ta­tive con­text, is that you want to make sure that you look at it, as the num­ber goes up and down that you’re also con­sid­er­ing that pop­u­la­tion might’ve been going up and down total, which will actu­al­ly like to assess the impact much better. 

So we’re see­ing here that in 2012 we had about 12% of peo­ple liv­ing in the United States that were not born here. And if you look at sort of the polit­i­cal cli­mate, there are def­i­nite­ly peo­ple that say, Oh yeah, there’s been an influx. There’s been a flood of immi­grants. They’re all com­ing here. And it’s been a ton. And the pre­vi­ous admin­is­tra­tion—” etc., etc., etc. 

So let’s take a look at what that looks like. Are we sud­den­ly in a flood of immi­gra­tion? Well, if we go back in time here a lit­tle bit, we can see that yeah, it’s been going up; it’s been up quite a bit. And now it goes to the 50s and the 60s, you can kind of see well, wait a minute there’s sort of a dip and now it’s going back up as we go back in time. And in 1910, we can see oh wait, it’s almost been 15% peo­ple in the United States that were not born here. 

And the con­text here is also that I see that that was actu­al­ly when a lot of peo­ple from Italy or Norway or Sweden came the United States. So we had a flood of immi­grants back then as well. And that helps me to actu­al­ly look at immi­gra­tion in much more con­text where I under­stand that if you talk about a flood, let’s take a look at actu­al­ly what flood looks like over time, so I can under­stand it bet­ter and see is it going up? Is that an unusu­al thing? Is that maybe some­thing that hap­pens in waves? And what are the poten­tial fac­tors that cause these waves? So that gives me a his­tor­i­cal con­text of the num­ber of when we say there are so many for­eign­ers liv­ing in the United States. Like myself. 

Let’s take a look at anoth­er exam­ple for con­text. This is a microsite that we designed for USA Facts last year for the midterm elec­tions. We call this a Voter Center. What it does is you can see in the bot­tom right here, it shows you stances of can­di­dates. These are things that can­di­dates have said on spe­cif­ic issues. And the Voter Center breaks down these issues into things that are rel­e­vant to active cit­i­zens mak­ing a decision—in this case voting—such as the econ­o­my, health­care, immi­gra­tion, defense. Things that real­ly are real­ly talked about a lot in the polit­i­cal agenda. 

And you can see the can­di­dates’ stances. And what USA Facts is doing is it’s jux­ta­pos­ing these stances with data—it’s not nec­es­sar­i­ly a fact-checking, because it’s not inter­pret­ing and it’s not biased, and not real­ly try­ing to take a stance. But it says if you lis­ten to this politi­cian, here’s the types of data that you should look at to estab­lish con­texts and under­stand­ing of when some­body says, in this case, that say the econ­o­my is doing bad. Or the econ­o­my is doing well. And it’s because of the admin­is­tra­tion, or not the admin­is­tra­tion. You need to look at a mul­ti­tude of fac­tors. You need to look at…for exam­ple a pop­u­lar one for econ­o­my is GDP—it’s the strength of the economy. 

But that’s real­ly only one fac­tor, so if we keep going, you can kind of see here, you can go down and see that there are oth­er fac­tors that you might want to con­sid­er. There’s like, how much pri­vate invest­ment is there? What might be the gold price? How many busi­ness­es are open­ing? How many busi­ness are clos­ing? What’s the employ­ment look­ing like? What’s the medi­an annu­al wave. Those are things that you would need to look at to con­sid­er if you want to have a clear image of how well the econ­o­my might be doing. 

Now you may look at this and say, Whoa. That’s kind of a lit­tle bit of a…you know, eye chart. It’s a lit­tle bit of a rough sort of slide here. There’s a lot of num­bers here. You’re throw­ing up a wall of num­bers.” At a design con­fer­ence; it’s a pret­ty bold move. 

I agree. There are a lot of num­bers here. And I do think that there’s a lot of com­plex­i­ty here. You’re absolute­ly right. And you sort of might be rub­bing your eyes a lit­tle bit. 

A photo of an old computer taking up an entire wall with a woman attending to it, contrasted with another of hands holding a cell phone

And I’m actu­al­ly gonna jump into some­thing that Bill brought up a lit­tle bit ear­ly this morn­ing, is we’ve done a great job I think as a dis­ci­pline, as inter­ac­tion design­ers, in tak­ing tech­nol­o­gy and mak­ing it easy to use. We went from hav­ing these machines—and Bill had a pic­ture of it as well—of these com­put­ers that would fill entire rooms or mul­ti­ple rooms, and you need­ed a physics degree or a math­e­mat­ics degree, and maybe even lat­er a com­put­er sci­ence degree, in order to oper­ate them and actu­al­ly derive val­ue out of them.

And I want to give cred­it to inter­ac­tion design to say we went from these these room-fill[ing] machines to some­thing that fits in the palm of your hand and for exam­ple allows you to man­age your men­tal health. We cre­at­ed We cre­at­ed these from these machines that worked for a very few and they could be used by very few peo­ple, to mak­ing them so easy to use that almost every­one was able to use them. And that’s a ter­rif­ic accom­plish­ment. So can sort of give our­selves a hand a lit­tle bit. This is great. 

However. There are cer­tain com­plex­i­ties that can’t be sim­pli­fied as easy as just reduc­ing them. We’ve done a great job at cre­at­ing plat­forms and tech­nolo­gies that allow us to com­mu­ni­cate with every­body, near­ly every­one, on the plan­et in a mat­ter of sec­onds very very easily. 

[Donald Trump, via Twitter]

But maybe some top­ics can not real­ly be dis­cussed with a suf­fi­cient amount of con­text that is nec­es­sary to ful­ly under­stand them in 140 char­ac­ters or slight­ly more. You need more con­text to under­stand them. They don’t fit. They are com­pli­cat­ed top­ics. They are not eas­i­ly simplified. 

We’ve made tech­nol­o­gy so easy to use that in cer­tain cas­es I can sit on my couch and get hours and hours of con­tent that’s specif­i­cal­ly tailored—that smart algo­rithms that are real­ly great will feed me end­less hours of con­tent that will fit my taste and help me enjoy my life. And in this case I don’t even have to do any­thing to get more of it. Some of the great­est inter­ac­tion design of all is sort of the inter­ac­tion design that does­n’t require the user to do any­thing, right? That’s great. Doing noth­ing is great. 

But in cer­tain cas­es, doing some­thing is actu­al­ly real­ly impor­tant. There are top­ics where we want peo­ple to take actions. We want peo­ple to take very very informed action. This map shows the edu­ca­tion­al attain­ment of women age 25 and old­er in some coun­tries in Africa, in this case specif­i­cal­ly Nigeria. So I look at this map and I can see alright, the edu­ca­tion­al attainment—this is the a num­ber of years of edu­ca­tion that women get on aver­age is about sev­en. I may know that in oth­er coun­tries such as the United States, Germany, the UK, that num­ber is about thir­teen on aver­age. So 

I can see now like well you know, Nigeria…it’s not doing so great in that regard. Okay. But I don’t actu­al­ly know what to do about the sit­u­a­tion. If I’m some­body who’s invest­ed in improv­ing the sit­u­a­tion I don’t real­ly know where to start and say, Well, I guess we’ve got to make edu­ca­tion better.” 

But if I add com­plex­i­ty to this real­iza­tion, I can now see that the prob­lem is actu­al­ly a lit­tle bit more nuanced. This is a map cre­at­ed by the Institute for Health Metrics and Evaluation. It’s part of the University of Washington. And what they do is they map health data specif­i­cal­ly but also oth­er fac­tors that fac­tor into health, such as edu­ca­tion, on a five by five-kilometer basis; these esti­mates do that. 

And it adds a lot of com­plex­i­ty. If you look at this now I see wow, there’s a lot going on here now. But what I also see is hm, the edu­ca­tion is actu­al­ly much worse in the north and it’s in the south, and in the south there are actu­al­ly pock­ets where it’s doing quite well. That’s interesting. 

So if I look at this I can now say if I am involved in this and I want to invest in edu­ca­tion in Nigeria, I might focus my efforts, my invest­ments, to the north. And I may actu­al­ly look at the south and see what’s going well there. And what are the oth­er fac­tors that actu­al­ly impact the sit­u­a­tion being the way it is in the north ver­sus the south. So the com­plex­i­ty here actu­al­ly inspires action. Whereas before, the sim­pli­fi­ca­tion sort of blan­kets the prob­lem where I don’t real­ly know where to start to address it. 

And so for us as inter­ac­tion design­ers, as we’re look­ing at prob­lems, as we’re being tasked with com­mu­ni­cat­ing infor­ma­tion, we should look at when some­body’s telling you, Well, we need to say that the quan­ti­ty of Blah ix X,’ we should strive to say we should help users, help our audi­ence under­stand the quan­ti­ta­tive of nature of X. How much is a lot of X, how much is a lit­tle X? We should also look at Understanding of well his­tor­i­cal­ly has X been a lot and has X been a lit­tle? We need the his­tor­i­cal con­text. And we also should look at all the fac­tors that mat­ter around X. What’s the rela­tion­al con­text that X exists in? And for those peo­ple in the audi­ence that’re sort of real­ly big on data, like­ly more than me because I’m just a design­er, I’m not mean­ing that these mean cor­re­lat­ed fac­tors. That does­n’t mean that there’s a sta­tis­ti­cal sig­nif­i­cance of say­ing well, if you look at the X then Y is always going up or down…” These are just fac­tors that mat­ter to what might dri­ve X, what X may dri­ve in regard… Those are the things to look at. Help users quan­ti­fy the num­bers, help uses see the num­bers in his­tor­i­cal con­text, and also show them all the fac­tors that exist that I should under­stand in order to form my opin­ion about what we should do. 

And I think as inter­ac­tion design­ers, we’ve focused so much on mak­ing things easy. I think we should use the tools… Surya talked about our tool­box­es this morn­ing. We should use our tool­box to make com­plex­i­ty under­stand­able. We need to use the tools at our dis­pos­al to build data lit­er­a­cy by show­ing the con­text that data exists in. Because with that data, and with con­text around the data, we’ll be able to build understanding—we’ll be able to under­stand com­plex issues, and I think we should all work togeth­er, strive, to build­ing under­stand­ing for every­one around us. No mat­ter what side of the issue they stand on, we need to talk about the same data. We need to talk about the same under­stand­ing that we have of the data. We need to all have the same con­text about the data. 

As inter­ac­tion design­ers we’re real­ly empow­ered to make it easy. We’ve man­aged to har­ness the com­plex­i­ty of tech­nol­o­gy and make it avail­able to any­one. Let’s all har­ness the pow­er of num­bers for us to dri­ve bet­ter deci­sion­mak­ing, and mak­ing those num­bers eas­i­er to under­stand so that we can inspire the right action that helps us live in a bet­ter world that is less biased, more knowl­edge­able, and yields out­comes for all. Thank you.

Further Reference

Session page

Help Support Open Transcripts

If you found this useful or interesting, please consider supporting the project monthly at Patreon or once via Cash App, or even just sharing the link. Thanks.