Transcripts For CSPAN2 Alberto Cairo How Charts Lie 20240713

Introduction to information graphics and visuals and the truthful art and his latest book, and he will be signing copies is called how charts lie. Also the tight northerly talk tonight. So welcome. Thank you. Thank you for being here tonight and thank you again to Northeastern University for having me. Believe this is the third time i make a presentation at this university so i almost feel like home. I would like to begin with a little housekeeping. The first thing is that i would like you to give an applause to the students who design this beautiful poster because this is not [applause] and their names and i will probably mispronounce theyre name, so, thank you so much for the these beautiful poster. I love it of like it even better than the book cover. More housekeeping. Youre interested in the talk, the talk you find it interesting on the concept im going to explain, find them enlightening or use. Theslides are not copyrighted and right after i finish giving the preparation, i will tweet out a link to the slides as a pdf file. So feel free to download the slides and use them as will in your classes or with your friends, family, your work in a company, you think the concepts im explaining may be useful for you, feel free to use the slides. A tweet i have ready to send right after i finish. During the talk i will explain why President Trump is over here hope to side of the image. So, the title of the book, the title of the caulk, how charts lie. I talk about i go beyond what the book describes and i repurpose some examples from the become. A little bit about the book. The title itself, condences part of what the book is about but the book is about much more than what the title says. For quite some tomb i entertain a much longer title that captures what the book is about a little bit closer, hour charts mislead, lie, deceive, how we lyle with them actively even to ourselves and this is a huge, huge problem, and i will address it later on during the talk. How we can book better chart readers and edkale of uses and can become better chart designers designers and why i we are in great need of using good charts that have meaningful conversations. Im a great believer in both the dangers of charts but also a great believer in the power of charts to illuminate great conversations. And this country and many other countries im familiar with, brazil, italy, spain, are in great need of good conversations in the future. So, i dont think i say anything new if i tell youve that charts are everywhere nowdays, so if youre have been paying addition in news media you may have noted an increase in the number of data organizations, maps, charts, diagrams, graphs. The rope why this is happening is through the analytics, the news media are observing that chart people like charts, people like seeing charts in the news that they consume. Whenever you put a chart or graph or map or a diagram in a story the engagement with the story increases and of you tweet out something or post something on facebook or instagram, if you add some sort of visual data graphic like a chart or photograph, the engagement with that post will increase as well. So theres power. Great ability, great news on one hand because im a journalist and i also teach how to design but at the same time i think we need to deal with what i call the systemic gap, and here what i mean. Throughout the years i have been on serving the sophistication of data organizations we see in news media is increasing and the sophistication of the dine sear design iris increasing the sophistication of the tools be. Thats also increasing so one hand in this spectrum of knowledge of this organization and statistickic cal thinking we have two groups that are growing apart, move eight part. Youve have what we could call the experts. Many of you are probably experts. If you are in this talk. People who know how to deal with numbers and visualize numbers, statisticians, but what worries me is what happens on the other end of the spectrum, with else necessary the public. Is the public brought touch speed with these new technologies . Does the public in general understand the graphs and maps and charts we see every day in the media, have the answer to that is no, or perhaps the answer is we could do a little bit better. One over the messages of the book and of the talk is that i think that we need when insay we i mean designer offed charts like myself perhaps we need to stop talking so much alves we do about the charts, the techniques, the design we use to create this organization and start thinking more about how to read them, and helping everybody else learn how to reason about the one reason which i say thesees is that throughout the year is have observed people in the fields of design and journalism, spread several myth it believe are detrimental to understanding Data Visualization, the everpresent picture is worth a thousand word. This is not true. Or this visualation is intuitive, map or chart and you assume everybody will fund it but a we understand its something we can all understand just in a quick look. We are without spending time understanding. The world of Business Analytics thinkday at that time should speak for itself. Show me the number. Data never speaks for itself. You need to make it talk, interpret it and then when you convey is interpretation of those data, or show, dont tell. Very common in the world of design. Usually joke that some of these myths can be made true if we append a second part of the sentence, append something else. A picture may be worth a thousand words if you know how to read it. If you dont know how to rate, then the picture is not worth a thousand words and this happens both with pictures, and happens with data solicitation which are abstract representations of data. Let me show you a picture, probably all of you will know how to read this picture. Probably many of you are familiar with it. It is a slogan. We all know how to read this thing and america runs on dunkin. Why die read this picture like this . Because i have been told how to read it. Nobody had told me how to read this picture i would read this like why is this person running away from the United States. That makes sense but its a little bit soso pictures can be greatly ambiguous. I have an anecdote the talk is an expansion of the book the first time my little daughter saw this picture she was four years old and i remember i i had the cup of coffee and she sad me, in a permanent british accent by the by a because she was watching a show called peppa pig and she stare at the picture and asked, daddy, why is that person waving from inside a toilet . And i realized that she was seeing a toilet, like a tilted companilet, and a person coming out of the toilet waving his or her hand for some reason. I tweeted these because i found it so funny and so revealing of how ambiguous images can be irtheyre not properly explained. I tweeted it out and someone two follows me said that doesnt look like a person waving from inside a toilet. Looks like a person from waving behind a martini glass. So pictures can be am big gauze and interpreted in multiple ways and if this happens with such a simple image, try to mam what may happen, hough ambiguous it can be to present to the public a complex Data Visualization without explaining. On my way here on the plane i saw a tweet promoting this graft. Its a tweet by an Organization Called data rapper which produces an excellent Data Visualization tool and have a web blog they pressured an article says the chart our developer finds another which country has booze more. You believe the russians drink the most alcohol, but this is actually not true. Take look at the data. They use dad to date to to show its not in the russians who consumer more alcohol per person per year and if you click on the link it takes you to the chart, showing the average Alcohol Consumption per year and person and you can see the russians are actually down here, the line has not varied that much. Actually the french who are consuming more liters of alcohol and that number has career diseased. Is the decreased. This chart good or bad . Never good or bad about whenever we see a chart it can take us a long way d. One of the first rule in becoming good readers of charts to Pay Attention to the source and to Pay Attention to the nuances in the data, never to take charts at face value. When you look at the source of the date to which is disclosed in in the article, in the nuances in the data and he limitations of the dat tack acknowledged in the article itself the designer offed the chart did a great job. When you read the article carefully. The article says we need to be careful with thieves data because this is only measuring Alcohol Consumption captured by government data. This is not consider the alcohol consumed out of the control hoff government, like alcohol you can produce at home which is tradition in many countries, like spain. So, the data has a lot of limitations and he cannot understand a chart like this just by staring at the chart. We need to look at the chart and then read the documentation, read the source and thats the first rule in becoming a good visualization reader and also the first rule in becoming a good visualization designer because this chart is basically just showing estimates and not capture entirety of reality, only a small portion of reality. Another feature of chartsful charts are never accurate representations of the world. Theyre limited representations of a portion of reality, and we need to understand their limitations. Opposite we do that, or of there as a consequence of these we should never assume that we can understand a chart at a quick glance. We should always stop and rate. Visualizations are maintain to to be seen and not read. Theyre visual arguments or arguments made visual and like any other kind of argue; you need to Pay Attention to it. Take for example what happens with hurricane maps. Ill talk about hurricane maps. This is a little about of a pet peeve of mind. This Hurricane Dorian from a month and a half oar or Something Like that. One of the latest forecasts her Hurricane Dorian. The National Hurricane center in miami, when they want to inform public of where a hurricane might go they use the cone graphic. The most widely used map to inform people where the hurricane may golf i present these example, presenting a lot of detail in the book so im not going to go interest a lot of detail over here. Just going to tell you beaut the limitations of the map and its great to illustrate one principle, if youre visualization design are, which is the followings. What you design is not what people see. What you design is that. That cone. That based on this experimental empirical evidence of observing how people read these map and actually interpret or misinterpret the map we know that many people who live in miami and the other places commonly affected by hurricanes, whenever they see that cone, they see Something Like that. They see a hurricane growing in size. Now, people are not stupid. People are not silly. Theres a very good reason why people see an area under threat when the say the cone. Thats a natural mapping. A very common way to represent areas under threat to use a different shade of color, right, in close with a line. It looks like the area may be under threat, but thats not true. What does the map represent . What does the cone represent . When you read the fine print, when you read the documentation of the National Hurricane center, they explain how to read the map correctly. What the map is representing is not an area under threat. Its a lot of probable positions of the center of the hunger in the next five days. Theyre telling you basically, the our most forecast is basically this line you have over here. We estimate this will be the path of the storm. However we cannot be sure. Therefore we are going to represent a certain level of uncertainty, why this is called the cone of uncertainty. The National Hurricane center collects tons of forecasts, mathematical models where the hurricane may go and these are lines. Then they come up with the open forecast which is base clay series of dots pinpointing where they estimate the hurricane may go and then surround the little dots with circled of increasing size, thats the level of uncertainty. We believe its going to be here five days from now but could be here, could be here, could be there could be anywhere and they connect the circled, and the result is the cope. I of you ever plan to move to florida this is how to read the map and its something you see a hurricane. What does that sunshine that means that areas that are out of the code may by affected by the storm because it the cone is only showing you the possible position of the center of the storm, and hurricanes are hundreds of miles wide. That mean s that even if you live, for example in this caught and South Carolina up here, you anyway still be affected by the storm down the road, five days from now. This map is also misinterpret but very important people. So this is the forecast on sunday, september 12019. You may notice that the day before, likely because it was not working properly this the forecast for august 30th. One day so it was predicted the storm would be basically one over the eastern coast of florida, but then in the morning, at 5 00 a. M. , the National Hurricane Center Published this map showing that dorian was predicted to move toward the atlantic ocean. And that same morning, a few hours later this map was published by the National Hurricane center, President Trump tweet, in addition to florida, South Carolina, north carolina, georgia, and alabama, will mostly be hit much harder than anticipated blah blah. Live ill saw that treat my brain malfunctioned because i had seen the map, i saw the 2010 and my reaction was basically has been, like, what his he talking about that what do you mean alabama. After a few days, im familiar with the case he insisted the tweet was right. But instead of correcting it which is what anybody would do, if i make a mistake, its veeries to misinterpret but the right action was to say i was wrong, was not very likely alabama would be hit so issuing a correction would have been the right reaction or the right response to this case but instead of correcting himself he doubled down, triple down, i was right, alabama was under threat, it was right and he even had a briefing in the oval office a few days later in which he showed a map not corresponding to the day of the tweet, map from a few days before help issued the tweet and thats the map he showed. Actually extended the cone of uncertainty with a sharpie and thats way we call this event sharpiegate. I looked into why did President Trump get this so wrong. This is not why he didnt correct he himself. That would have been the right reaction, i believe. But why get it wrong . It was because he exposed on the days heading into september 1st to tones of maps tops of maps showing one of them, one from august 30th, the 30th, few days before the tweet, showed there was a low probability of alabama being affected by the storm. It was a very low probability, 10 . About a few days before the tweet the was a map showing that and another map, up earlier, one of the first maps produced by the water service, i believe, showed that some of the forecast models run over alabama. Most of them showing the hurricane was going to go over florida, later forecast showed but few of them that this wily touched alabama. The problem ways the map what personned much earlier than the tweet forecasted the hurricane moving to the atlantic ocean. Id like to yell beaut that map not in the match itself but the caption of the match. The caption is perfect to illustrate the problems were dealing with today. The caption says, if anything on this graphic causes confusion, ignore the entire product. I absolutely love these captions. I love this so much i even consider it as the title of my next book, the followup to high charts lie, because it captures perfect live the powers and also the dangers of Data Visualization, sometimes it beck extremely confusing and hurricane forecast maps can be very confusing. Willing to give everybody and anybody the benefit of the doubt. I could have made that mistake miss, tweeting alabama in the tweet, but you need to correct yourself. So what is going on sneer here comps nerdy part. The extra dillsal modeled put the designer first. When the designer, anybody who produces the visualization is designing a chart or map or a graphic or whatever you have a certain mental model our the visualization will work and what that visualization is representing, what you want to show. But also i said before, what you show is not what people see. Why not . Because when a reader comps to your graphic that is corrected rare. Colorados with completely different kinds of meant models. Maybe not the same level of knocking you do. When theres a mismatch between the enemy model the key cyberis using the mental models the reader is using, thats when mr. Understanding happen and theres a third component which is the role of mediators or enter mideast areas. They could be journalists, theres a very good reason why the cop of uncertainty is so widely misunderstood. Watch tv next time that you see a hurricane coming and notice how tv journalists explain the cone of uncertainty. They usually get it wrong. They usually explain it wrong. Usually complaint as an area under threat. Those are the mediators. The mediator have a responsibility to complain these visualization as well in order for the public to understand better. Whoer the mediators in the case of sharpiegate, the meetadors will be the people briefing President Trump about the maps. Another reason why he got that wrong and sent the tweet is because that morning, september 1st, sunday, he didnt receive his briefing. Just remembered the maps from days before, and made the tweet base on the maps because he had not received the briefing from hi people. What can we do . We can do self things. Obviously one of the best things is to Pay Attention to any graphic. But we need to assume that in any situation or many, many situations, we can assume one of these roles. The designer, the reader or the mediator, but in many cases well assume more than one role. And i will get to what i mean by these. Depending on the role we are assuming when dealing with a chart row can remind mouthses or ask yourself certain questions. Were the designer we asking ourselves how can we help our readerscrest the right mental models and you can use convention or explain the graphic, explain how to read the graphic. If you are mediator, if you a journalist or a tv newscaster or whatever, how can we make this translation . A mediator is a translator between the complexity of the time and if core 0 read youre, how to educate yourself to become more graphic acy. It as term invented in 60s, to refer to graphickiccal literacy, and unfortunately theres a huge, huge lack of knowledge of these kind of literacy. Has written tons of booked about maps, how to lie we maps an absolute wonderful book about maps. The bees introductions to cartography i havent rate and he has another book and says nowdays in order for a person to consider herself or himself educated citizen in the a republic such as the United States we need be literate, need to know how to read and write but thorne half. He need newman was si which is not exactly numeralsy is not actually scientific thinking. Its more like a sixth sense you can develop by studying that , statistics, mathematics, in order to deal the sixth sense in the back of your brain, like an alarm that will start ringing whenever you see a number that looks dubious or a chart that looks strange, the will ring in the back of your brain, prompting you to look deeper into the number so see what lies behind the number. And its an extension of numerousy or supplement we have graphicasy, the alto interpret and it seems we still have a lot of work to do , book in 2014 in the Pew Research Center conducted a survey in which they ask a sample of 1,000 people, what do you see is in chart . The chart is showing the association or correlation if you previous to doc but the lineal mod association between average sugar consumption per cap a to and country and then the average number of the teague per person. Theres good news and bad news about the results of the survey. The bad news and good news is that 63 of people in the survey could read this graphic well at the gramatticcal level. What do i mean . They could interpret that each one of those is a country. Could interpret the meant the position in the herossal axis means something and the position on the vertical axis meant something else. Thats gramatticcal level. This is great news because as a journalist i can tell you if we had conduct third survey 20 years ago the permanent would be mump lower. Why . Because nowhere to be seep in newspapers 20 or 25 years ago. The first cat to plotsre purchased do in the United States, 10 or 15 years ago. Therefore, this is just a guess, that percentage is ooh higher today because more reader have been exposes the these charts throughout the years the bad news is we still have a third of people who cannot read this graphic even at the gram matical level. Some people interpret the chart of showing change over time. Why . I dont know but they read that. They read that. The graphic is confusing. Didnt have the right level, the right level of literacy or graphic late rays to read this because this is a type of data that has been around 100 years. It was studied correlation and association and regression and stuff. So what can we do . We can show and we can tell and we can explain. One person that i believe we can know looking to for inspiration how to present data but also how to explain data and the same time how to explain the graphics that are represented these data is the professor who passed a couple of years ago, professor hans was a professor of International Health in sweden. Became visit in 2006 because of a ted talk in which he made data sane, made date to talk and speak. You of that the opportunity to visit his website because his website collects a most of the talks he gave before he passed. He also published a book called facts which is quite interesting. This is just a few images of the little bit of footage of a documentary me made a fewyearold later called the joy of stats in which he presented a chart before he shows the data is explaining their vertical axis, the horizontal axe sis. Means income. The vertical means years. Ill show you this quadrant corresponds to countries with a health goody and live long. Now let me show you where the countries are in the 19th 19th century. Each one of these dots represents a country. A way he was explaining they data. By explaining this graphic the was u. S. Knot just explaining the con at any time of the graphic it. He was increase thing if level 0 graphic of his audience by explaining how to read that graphic. He was a very inspiring figure. One of the most influential. The second matter, the second recommendation is to embrace complexity. This is for the journalists in the room. I know because im working in journalism for many years if you ask journalist what our main task is, some of us would say, we want to simplify matters. We want to simplify matters for our readers because we want to present information thats understandable by everybody. To simplify is very dangerous. I prefer to use the verb to clarify, borrowing from my friend nigel holmes. What is the difference between those two words . Those two words might put you in a different frame of mind. When you are in a frame of mind of supplication, all you care about is about reducing the amount of data you present to the public. We should not overwhelm people with numbers and tons of numbers that are needless to understand the story but sometimes that drive the impulse to simplify can lead us to oversimplify matters. Sometimes we need to increase the amount of data we show in order to clarify a story. Take a look at this chart. Theres nothing wrong with this chart per se but the chart can be misinterpreted very easily. In the past two or three years if we could extend this line up to the present that line keeps increasing. I know because i heard this that many people will describe the chart as the United States is becoming a much more dangerous country. And the descriptions verbal descriptions of the chart we see every day can greatly biased our perception of the chart. The chart is not showing that the United States is becoming a acountry. Most places have the United States are very safe. If we could plot every single town, city, neighborhood, whatever unitive geography you want to use, on the chart on the vertical axis, most places in the United States and that would be down there. Around the National Rate or below the National Rate. Imagine each corresponding to a city town or place in the United States. Whats going on over here . Extreme values. There are certain cases or certain places in the United States that have become so dangerous in the past five years or so, the murder rates have increased so much in these places that they sort of distort the National Rate. They are like a magnet pulling the lineup. This is a perfect example to illustrate the difference between to simplify and to clarify. If you only show me the National Rate if you only show me that National Rate is a journalist publishing a story about these newspaper you are assuming that i already know that there are outliers that might be distorting the National Rate. That simplification may be obscuring the information for me a reader who doesnt know that much about these issues but instead of doing that, if you show me the chart of the National Rate is also an explanation of this extreme values that might be changing the pattern or trend of the data by distorting the liner moving line up or down the extreme values these places that have become so dangerous in the past few years now you are clarifying by increasing the amount of data you are showing therefore we need to put ourselves in a frame of mind of to clarify not to simplify. What else can we do . If you are in not training in statistics or data science. This is the case of most journalists, it can take a long way to read a little bit about those matters. Fortunately nowadays there is plenty of Popular Science books that deal with the main ways that statistics may lie to us and how to interpret them correctly. Recently we had this book that was going to like a month ago the art of statistics. We also have the classic negative statistics a great intro to the most elementary message in statistics. Like some of you in the room who are not formally trained in numbers to see if these books can help us bring ourselves up to speed with all these matters. We can recommend them to other people as well. Theres a difference between supplication of the data and a clarification of the data. An understanding how to deal with numeracy. Recommendation number three out of four, there will be four recommendations over here would be to become more mindful about our bias. Both ideological bias and cognitive bias. The most thing with cognitive bias where you would be the confirmation biased how prone we all are to project to project what we what we like to see into the charts and stories we see everyday. It happens to all of us. We can control ourselves we can become a little bit more mindful about the tricks that the brain plays and consciously in order to deceive us. Money showing example of the many how the book touches. So i show you the chart and say cycling of the chart. As a journalist i can tell you that someone like me will describe the content of this chart this way and i know this because i have done this myself but most of mistakes i describe in the book are mistakes ive made myself a step many people will read these chart. About mental descriptions. Many people will describe the chart the more we smoke the longer believe. That chart cannot be described that way. That doesnt mean that these two things are connected to each other. In other levels of aggregation, dissociation is positive correlation we see in there might be disappear. Thats the right way to explain these charts. One of the things that we need to remind ourselves whenever we say chart hits another ruled the book that i did put to the top. Another one of these that can help us become better chart readers is that a chart shows only what it shows and nothing else. Everything else you see in the chart happens in the interaction between the chart on your brain. This chart is not showing that the more we smoke, the longer we live or the more i smoked the longer i live. Because this is only showing data aggregated at the national level. Its not showing individual person by person data. In fact, this is a great example to explain to people certain paradoxes and certain biases and problems whatever we interpret data. For example, abor simpsons paradox or abheres what i mean, if we start desegregating the data. If i start separating the countries by income levels, take a look at what happens. There is no association anymore. Now im grouping the countries by income levels. The strong positive a cessation that we had before becomes much fuzzier. Because we are dealing with several underlined lurking variables we are not taking into account. For example, well. The old saying correlation is not abwho wealth is available we are not contemplating. The wealthier people are, the wealthier country, the more cigarettes people may consume therefore the Life Expectancy might get lower but at the same time, people who live in wealthier countries tend to live in safer environments and also have access to better healthcare. Therefore, the Life Expectancy you may lose by increasing the cigarette consumption may get balanced out by the good healthcare and the safer environments. Correlation and concession isnt the whole story. What about abparadoxes and ecological policy. A pattern that you see in a certain level of aggregation national level, country by country more cigarettes, more Life Expectancy but if you go to a lower level of aggregation. Person by person data, the association positive originally might become negative the more cigarettes you consume, the short tour you will live. This has happened this happens in this particular case. When we go to the individual level, person by person data the positive association we had before becomes a negative association. If you take a look at data person by person we know already that consuming, consuming more cigarettes in general leads to shorter life in general. Cigarette consumption, short lives. We dont know that if we take a data at the lashthe national le. Whoi want to know whether smoking is good for me. If you dont know how to read a chart like this, the original chart will mislead you. It will reassure your opinion or your decision to smoke to keep smoking. How to deal with these im going to recommend tons of books today. Sorry about that. I love reading and i believe reading is the best path to knowledge. There are plenty of books nowadays we can consult to learn about our Cognitive Biases and perhaps how to curb them. These books talk about how the brain plays tricks on us but at the same time, also gives some hints of how you can curb that by becoming more mindful by becoming more attentive to what happens unconsciously inside our brains. Mindfulness. Paying attention to what happens inside our brains. We have the classic mistakes made but not by me. Like carl athis is an excellent introduction to Cognitive Biases. The second one is one of my favorite the enigma of reason which basically explains that the theory that our intelligence are cognitive capacity to understand the world to the process of information it didnt really evolve to help discover how the world really works. It evolves to persuade other people and to persuade ourselves about the rightness of our own positions. In order to attract people to our own groups and own tribes intelligence apparently according to these authors evolves as a mechanism to convince and persuade. Only that we can coopt those mechanisms to understand that role as well. The knowledge illusion explained that knowledge doesnt happen in individual brains. Knowledge is distributed therefore we need to understand that process. In learning a little bit about the church of mindfulness and attentiveness and paying attention to all these issues inside of our brains. It will help us be more ethical. Its a final part of the presentation. I would like to circle back to the diagram actually before with the triangle between the designer, the reader or the user and the mediator. And go back to the theme explain at the beginning of the talk which is nowadays very often we all assume these three roles depending on the situation. As i said before, the association gives us a lot of power. As a classic movie says with great power comes great responsibility. We need to be responsible of those tools. If you are a designer of ab you are responsible to help people understand the data you are presenting. If you are a mediator and all of us are mediator was now day because we all have social media. Therefore we are all journalists and some sort. We distribute the information we are all mediators every day. We have responsibility to verify what we are presenting to our network. As readers we also have responsibilities responsibility to be attentive to the information we see every day. We need to cultivate a culture of attention, a culture of attention. There is a wonderful book a final book i would like to recommend today by professor from this university. The social fact, ive been reading in the past months or so. Read it. He really explains these network of connections. In past models of information and distribution we have a ab of information edited rude deceiver of aba receiver of information. Today we all contribute. Its a network. We repurpose it we are all part of this network and we are all responsible for making this network work correctly. This is an ethical responsibility. One of the classic books in the history of journalism says the essence of journalism is a discipline of verification. Its from the elements of journalism where the classic introductions to the world of journalism. The first time i read that sentence something popped up in my brain, its not just the essence of journalism. That is a discipline of verification. All of us are nowadays in social media we are all distributors of information. I would say the essence of citizenship in an informed republic and a more informed democracy is to be a discipline of verification. We are all responsible to create a better informational environment. Im going to leave you with the last message today which basically repeats what i just said, we need better conversations this goes back to the beginning of the presentation and in order to and form these conversations we need to take into account all the roles i explained before. We all have an ethical response ability to be better designers, better readers and better mediators. Which is the role we assume more often. Thank you so much. Thats all for today. [applause] absolutely. Im not in a rush. Hello. When you are talking about biases. Another heuristic came to mind the availability heuristic. Which i think plays a powerful role in how we feel our confirmation biases. We see so many newspapers reporting murders they dont report the people who were murdered and so all of a sudden you think that the world is a much more dangerous place that it actually is. The abecause we see more often. Yes so we recall that event more often. Its more a the social fax network has made it even more verbose and robust. I completely agree. Obviously in just one hour i couldnt talk about all the biases. Theres hundreds of them. They connect to each other the reason why i came back to this slide over here because the abis a huge one its very well covered in the first book. Its connected to the confirmation bias. If we dont Pay Attention, if we respond to the message we receive every day like automatically, we dont stop for a second and think about my being biased by all these news about murders and violence . Is it really a representative of whats happening everywhere in the country . Ask a question that you need to force your brain to ask itself. Then look for more information. I do have a question for you. At the beginning you were talking about the ambiguity of visuals. You make an odd fan comment i was wondering if you think ambiguity scales with complexity. Literally or nonliterally . I dont think what if we know if we can really measure that but yes, the more data and information you put in a chart unless you explain it, the more ambiguity, the more people will tend to see what they want to see. One of the ways i explained that i love abbut they can be greatly often misinterpreted forcing you to see coal station is versus just association. If you have several scatterplots, thats scales you have the opportunity to see multiple causation links between what youre representing. Theres a reason why put so much emphasis on the role of the presenter. The data is showing at, however, be careful not to see all these things. You can reduce ambiguity through certain principles of design but you can reduce ambiguity by having that person explaining the graphic to people. I believe we should amore often we should show and tell people rather than showing and just walking away from the chart. We call that in the world of visualization the role of words spoken and written the annotation layer of this. Sometime we dont pay enough attention to and we should. I have a question about your mentioning how we often say that a picture is worth a thousand words in a visualization should be simple and you should be able to look at it and know exactly what youre thinking, i hear that a lot, averted a lot in newsrooms and have tried to push against it. Ive been trying to explain, if we allow readers to read however many hundred words and get to spend time with that, we should also give them the chance to spend that much time with the data and the visualization. The pushback i often get is that the words and the story are important and that the visualization is just eye candy. How do you push against that or how would you suggest to people in here to push against that idea that the real story are the real journalism is whats written by a narrative writer and that Everything Else is sort of in service of that. Complementary to that. Use examples. Sometimes it works and sometimes it doesnt work but it has worked for me in several cases in the past. If you get the pushback, ask that person, who are your favorite sources of information . What are your favorite newspapers . Favorite magazines, favorite websites . They usually would say the New York Times. Whats the Gold Standard of investigative reporting . See how they work. They put the words and the visuals at the same level of importance. The visuals dont complement or supplement the words the same way the words dont supplement the visuals necessarily. Sometimes they emphasize more visuals and then make a decision based on discussions about how to shape the stories. One of the Gold Standard in investigative reporting but also one of the Gold Standards in the use of visualization. The followup question is, do you want to be more like the New York Times republican, this what they do. Sometimes the argument works, sometimes it doesnt. We need to keep fighting. These books are like weapons. Take a look at the chart. You find it illuminating but notice is not simplifying information. Its actually very rich in the amount of data. Its a longterm fight. We all get that pushback. You made a statement that citizenship is verification a not only but part of it. Being able to verify information that we dont have primary source documents i rely on you and your reporting on hurricanes in florida and twitter. You kind of already answered it previously but part of that is like a bit of privilege for having understanding of Data Literacy and graphic literacy. Where do you draw the line of who i trust and how i trust them and also in the circumstance that i dont have this knowledge of how to interpret graphs . Thats an excellent question. Its one of the things i deal with in the book. On one hand, i have two or three answers to that question. We try to phrase this in my mind. The first thing is that we can still verify. Even if you dont have, abi read the news in spanish radio for three months during an internship. I can still read the description of a study in the narrative description. I can see take a look at a chart and say that the chart is showing murder rate i can still go to the source and read what they mean by murder rate. I can understand that even without a deep knowledge in statistics. Thats what i mean by verification. I dont mean that you need to go extremely deeply into the data or understanding deep levels. Its very elementary. Very shallow level of verification that only takes 30 seconds, one minute, make sure that what the chart says that its measuring is actually what its being measured. That can take you a long way to avoid many mistakes. Im basically spreading charts that are dubious. One of the rules i have in the book is that if a chart doesnt disclose where the data comes from, if the source is not disclosed, distrust that short period. Thats the first rule. On one hand i think anybody can become more used to, i like this chart i have a couple examples in the book of chart i retweeted myself mindlessly and then thought twice, like wait let me look at the primary source. I have an example of how to do that because the book is written in a way that can be adopted by High School Teachers to teach these very elementary techniques to high schoolers. So dont retweet something mindlessly come up with something mindlessly in instagram, open the primary source, read it for a minute and ive done that it will reassure yourself. It looks safe now i will post it online. We can develop this it will take very long it will take a lot of education but we can do it. Thats on one hand. The other hand is that we cannot do this in every single case. It takes one minute and start adding up. We can develop a healthy media diet. I have recommendations, i dont know if they are appropriate for everybody or applicable to everybody, of how i did that myself is that you cannot do that immediately. You need to do it in the long term. Following news media from all over the place, way whether these media sources but you are getting information whether they correct themselves and how often the issue of corrections because the amount of corrections on Media Organization issues actually correlates with equality. That Media Organization, thats a good proxy variable then if you start noticing a pattern of deceit, just push the organization aside. Thats how we can develop a healthy media diet. The same way you Pay Attention to what you put in your body through your mouth, what you eat, we also need to be a little bit more mindful of what we put inside our brains. We need to develop we need to cultivate. That sort of media diet. I know im talking from the point of view of a person privilege. But i think the basic techniques of how to develop healthy media diet can be talk to anybody. It can be taught to middle schoolers and high schoolers. Im optimistic about all these. We have several questions over there. Over there. Thank you very much for the very interesting talk. He mentioned new technologies and how they change the way we communicate. The use of more charts and journalism, social media, but we are still using trying to replicate paper and ink. So the slides are flat, New York Times is still trying to look at the New York Times on the screen. We still read books. My question is, what do you think is the future of media and have you seen any way of any sort of media that made you feel excited about new ways of communicating both for journalists and researchers. I can only talk about visualization. Graphics are all about the only thing i could talk about with any sort of authority. Lemme go back to this for one of the first slides i showed you in the presentation because this applies and relates to your question. On one hand, im seeing new technology that looks very promising to me. Im a little skeptical, useful they might be. What im most excited about in terms of those technologies is there applications for pictorial representations. Imagine you map a highresolution model of the human heart of a patient a doctor and he put on your glasses and you can walk inside the heart and see the heart from the inside. Thats based on a real example i saw a while ago. The applications of this technology that are extremely promising. However, as i mentioned when i was presenting this. I think these technologies apply to these ends of the spectrum. Im not that worried about people like you. You are on this end of the spectrum. For people like these, for friends, family, my own dad, they need to bring themselves up to speed with the traditional 2d graphics. Those are still extremely useful. We need to help educate others and ourselves. We can do both we can multitask. Developing new technologies but at the same time strengthening the knowledge and the understanding common understanding and societal understanding of the traditional technologies to represent information. Otherwise this gap is going to grow wider and wider and we cannot allow that to happen. Thank you for your talk. My question is about the clarification and simplification from your talk. The more i think about clarification it seems to me that you are adding more data to your graphs come in that case, youre probably increasing the complexity. If you think about a reader who is new to charts, are you in a way kind of confusing them more and telling them 20 past they can follow . Versus simplification is like show the part it wasnt true that the crime rate was rising but may be a simple annotation would have been better than clarification. But thats exactly what i met. You can still show the chart but you need to disclose the existence of the outliers. If possible you could add secondary charts showing where those exceptions are. That doesnt add a lot of complexity to the chart it could greatly clarify the information. When i was explaining that i shouldve explained how i spent a little more time than that. I dont mean in every single case we need to increase the amount of data. We need to emphasize certain aspects of the chart in order to make it clear. However, there is a common saying traditionally ab talking about science. Apparently einstein once said everything should be made as simple as possible but not simpler. Thats the clarification. You can certainly reduce the amount of data. To simplify but there is a principal in there that if you go across that it will be oversimplified. Sometimes you need to increase, the amount of information you provide. In order to understand the story or help people understand the story. How can you decide . I imagine spectrum between showing every single data point all the complexities of the data and variables showing a little data in the other kinds of aggregations. So theres a spectrum of should you be here or here. That a function of a ton of different factors. What the story is in the nature of the data. And what is the right place . Its a decision you need to make weighing all these factors. With your own understanding of the data. In your own understanding of the public. One of the things i didnt include in this version of the talk is to test the graphics more often. To show before we publish our graphics out there actually show that graphic to friends, families or whatever. Have them read the graphic for five minutes come back and asked them what did you learn. What did you get from this graphic and have them talk about that. Its not a very scientific way of understanding misunderstand the graphic it can be greatly illuminating. Very illuminating to have people talk about the graphic. It is one reason why you put so much emphasis on the cone of uncertainty. Because i do the exercise myself the first time i saw the cone of uncertainty i described it to myself wrong. We conducted several focus groups in miami and a Research Program part of which we show the maps we ask people to talk about them openly and freely. Its a very funny one. Im not going to make fun of the participants of the studies but its another mr. a misinterpretation of this. Its not a continuous. We are one area shaded in one area darted. Whats the difference between those . Theres not a qualitative difference between shaded area and dotted area. The shaded area is a forecast in the next three days. The dotted area is a forecast in the abwhen people see the graphic, this is a little bit isolated in one corner. When he asked people to describe what you see in that dotted area some people would tell you thats probably rain. Thats a natural mapping. People are not stupid. People will Pay Attention it all depends on how you show it. I would be on your question but these are things that i mentioned. I forgot to mention before the talk. I think they are important. We have more questions over here. How would you redesign that . Who is talking . Over here. How would you redesign that if there were so many interpretations . Asked me one year from now. I can tell you, im going to talk against designers now. Designers have very strong opinions and probably his desires and a regular life these people abif i were myself i would desire this way. Your assumptions are wrong. All of them. I try them all. We might come up with different solutions. I dont know what the outcome will be. What i can tell you is that i have nothing against these. I saw people, friends of mine who didnt know how to read the graphic before that after reading the piece, now they know how to read graphic because going to explain it to them. Understanding is not a function of the sinus function of the explanation we attach to the designs we present to the public. Im a huge believer in putting a human face to the data we were present. So the role of the mediator. I dont have an answer but i answered it anyway. More questions . I think i have the mic next. Hi. You touched on this briefly with the previous question and the question on verification. Emily interested to hear your thoughts on who bears the responsibility. Every time the term literacy comes up those contentious arguments on, you can run into the problem of creating an ivory tower when you have this group of experts. Or maybe condescending by claiming you need to educate people. Do your slides or get the initial impression thats maybe the role of the mediator or designer to obtain or create graphic as he but as you mentioned, these roles are blending not everyone is only one role. Who do you think has the responsibility to attain it or produce it . When we go back to the slide in which we have the triangle. Primary responsibility is in the part of the designer. In part of the communicator. A huge responsibility on the part of mediator the translators people who get information from experts and presented to the public. And the guy with the sharpie, that person also has a responsibility. Im a great believer on individual responsibility. One of the things, this is just on the side, one of the things that drives me crazy about current discussions on the role of technology in misleading people is that all the brain is put on the platform. Facebook has a huge responsibility, twitter has a huge responsibility, they are on the side so they have primary responsibility. Sometimes we forget sometimes its us. Its us falling victim to our own biases. I dont mind being educated, i love being educated. Show me, teach me, explained to me, i love that. Thats the reason we read books, to be educated. Theres nothing condescending and using that language, let me educate you about how to read charts. Its not condescending. I know about chrtas, you probably educate me about astrophysics or whatever. Her cognitive bias. Im not an expert in any of that. I have many friends who have educated me and using numbers. They were not condescending at all. They were working as mentors. We can mentor each other. We can create in this network of help and talk about mutual mentorship. Sharing knowledge. Knowledge doesnt reside in signal brains. Im rambling a little bit. We have more questions. Can you speak a bit about accessibility in visualizing data and how we visualize for people who either cannot see or see a not with any degree of authority about that. Will that help increase accessibility . Maybe. Hi. Where are you . Im right here. He mentioned misunderstanding sometime stems from the readers having a different mental models and designers. And wondering if we should add it to the responsibilities as designers and journalists understanding other mental models and designing for other mental models and that may include also how things look. Kind of to bridge polarization. It all comes down to what i mentioned before about testing. Presenting data to people, see how people react, see how people read it, see what mental models apply in terms of understanding the graphic and the contents of the graphic that we present to them. Learning a little bit about human factors, it can take a long way. Im not an expert in any of that. Learning a little bit about the literature of polarization. There are plenty of books published recently about societal polarization. Many of the books i mentioned before, what they say is that the traditional model of human understanding, by that i mean we see data and have the data and process the data and base opinions on the data we see in the evidence we see. Its actually not right, what we do is the opposite. We first form an opinion emotionally and we tried to recruit the data to confirm that opinion. Thats the confirmation bias. Its a mental process we can understand and predict and prepare for. Whenever we design a graphic. I know it sounds very vague but it all comes down to reading, learning, studying and then testing presenting the graphics to people. And see how they react. Thank you. This is a really nice talk. In globalizing interconnected world. Have you come across annotation clarification that my work very well for an audience in america or audience a but then when crossing to audience be suddenly muddies it. Culturally . Yes. And if so, what are some examples of that and how do you fight against it . Is there any emerging techniques to create visualization that are more toward the global audience or toward the Global Citizen of today . Unfortunately i dont have an answer to that. You notate a graphic obviously uses in the native language that people speak. Each language has its own nuances. For example, i speak english, portuguese and spanish, portuguese and spanish are quite similar but if you write in portuguese the same way you write in spanish, nobody will understand you. You have to use the language the symbols and the words. So im not going to answer the question because i dont have an answer to that, however, i would say that one of the Research Agendas that i hope that somebody will adopt at some point in the world of visualization is crosscultural visualization or comparative culture whether people coming from different languages different cultures and environments whether the rules or principles of this are universal or whether they are shaped by the culture that people come from. I tend to lean in both ways. Its so preliminary, i tend to believe there are certain principles that are may be universal because theyre based on how the human brain evolved. But maybe other principles of visualization that might be dependent on culture. Theres a reason why visualizations look very different in different countries. Theres a culture that might affect the way people design and interpret visualizations. Sorry for not being able to answer the question. Hello. Im in the back. Im really struck by and i want to thank you for mentioning the divide between people who understand how to read charts and the need for education. I know that you kind of track the scientific Visualization Community pretty well. It seems like we are still trying to figure out how pie charts work. Theres papers about areas and i wonder what this lowlevel research has in this education and figuring out how people perceive things. How do you see it all coming together to actually bridge us . That connects to the answer to the question you asked before about abim super excited about highend innovation. New techniques, but im also super excited about basic research. To give you an example, very elementary example. Theres like a rule setting in visualization which is basically if you going to design a bar graph, dont crop the y axis. Its a rule that is still i read about in the book. Because i say, the encoding going to the height of the bar if you club the y axis is going to destroy the bar. However, there is a caveat in there. We dont really know yet if most people when they read a bar graph they focus or extract the meaning of the chart by focusing on the shear area of each bar or if they focus on the upper edge of each one of the bars. If they focus just on the upper edge, there is no really difference between the y axis of bar chart and truncating the y axis of a line chart. Which is sort of considered a rule and Visualization Community. It makes sense not to truncate the y axis of a bar graph because the way to represent it is the height of the bar. D encoding where we call ab theres a huge space in terms of vision science, cognitive science to explore how visualizations are read. Im super excited about that because thats what informs the principles or rules of Data Visualization used every day. Theres still lots to be discovered. Unfortunately sometimes people who write about visualization, including myself, sometimes we sound too certain about what we know and sometimes we know much less than it seems. All of us i think. We need to be a little bit less certain about our opinions. Two more questions. We had a question a few minutes ago about differences in cultures. I would go under if you could comment on difference. Huge differences. And there should be. Theres nothing wrong with those differences. Visualizations let me rephrase that, it comes to the reader, whos good to be the reader thats always the key question. When a scientist produces visualization it usually is because the scientists know shes gonna talk to other scientists. She designed the graphics in a certain way to present the graphic to the community that is supposed to share a Common Knowledge base. The cultural journalist doesnt make that assumption or should it make that assumption because we usually create graphics that are to be consumed by people coming from all over the educational spectrum. We need to make other assumptions. People who come from Business Analytics they make assumptions about presenting it internally in their companies they assume a certain level of basic knowledge about the internal functioning of the companies. Certainly there are huge differences cultural wise. The problems arise when you have your frame of mind adapted to a certain culture and you come out of that culture and tried to have graphics in a completely different environment and do it wrong. At the university i sometimes do workshops for communicators, journalists, graphic designers, market aand sometimes workshops for scientists. What i emphasize in those workshops is completely different. When i talked to marketing, journalism, emphasize the data understanding part of the process. The principles of Data Visualization based on the data. When i present virtualization classics to science i usually talk a little bit more about the storytelling or the narrative part or annotation part. How important it is not to assume people know as much as you do about the conduct of the graphic because you need to overcome that what some people call the curse of knowledge which is something that ails all of us. I have two questions, first is where you get inspiration and what your three favorite visualization tools. One of the ways i learned i need to copy them i dont mean plagiarizing, i mean getting inspiration. One of the best things about the visualization communities they are very welcoming. You have on one hand Data Visualization society which is something you can join for free and pause abother Communities Just makeover monday, which one data set over week and they try to redesign it. Its great. They are very welcoming. I get inspiration from all of them. Its difficult for me to pinpoint specific designers nowadays. That i really admire because there are so many. Im a huge fan of the New York Times. Which is the usual suspect. The huge fan of probiblical. I think republican is a great data journalism. I use whatever to get the job done. I you quantum gis. Which is a tool for cup tropict so many. Its difficult for me to say. You asked for my favorite. Sometimes people ask for the best tools. The best tools would ever get the job done. Whatever works for you go for it. I use the art programming language but some students of mine preferred abokay fine. I would never say dont use it. Only with a very low voice. Python is fantastic. If python works for you, go for it. If you prefer to use whatever other tool. We have a question over here. Last one. Before i forget, if you want to see tools come in my website, before i forget let me tweet these out because i have these tweets ready. Im going to tweet out these slides. Twitter. Com alberto cairo. There are a ton of tutorials about excel, road graph, flourish. About insight, which is a designer tool about what he mentions. Inside is a tool for data analytics. A free open source. Its for quick dirty analysis. Do you think we need a tool to represent abin the visualization . He mentioned the big gap we need to solve is to do to athe public. I know the correlation doesnt represent aat the college level. Also most public readers i know, they care more about causation rather than correlation. [indiscernable] do you think we need to design some new tool to represent this . Perhaps. I have no idea how i would represent that. Perhaps a qualitative abstract diagram with arrows. One of my examples is to how to represent different uncertainty. Thank you very much. [applause] the new cspan online store now has booktv products. Go to cspan. Org store to check it out. I focused on using miss skills to advance opportunities for people with disabilities. Why did you choose that type of law . I was born deafblind, most of our world is designed for people who can see and hear, and when