What Data Visualization Means for Data Literacy

Adel Nehme: Hello everyone. This is Adele data science, educator, and evangelist at DataCamp. This is week three of our data literacy month special at DataCamp. And this week is all dedicated to data visualization and data storytelling. And there's no better person to talk about these topics with than Andy Cotgreave. He is co-author of The Big Book of Dashboards and Senior Data Evangelist at Salesforce in Tableau.

He's the host of If Data Could Talk cohost of ChartChat, and columnist for the Information Age. He's also on the 2021 data IQ, the top 100 most influential people in data. With over 15 years of experience in the industry, he has inspired thousands of people with technical advice and ideas on how to identify trends in visual analytics and develop their own data-discovery skills.

Throughout the episode, we speak about why data visualization skills are so important. How data visualization skills can drive organizational data literacy, best practices for visual storytelling, and much more. If you enjoy this episode, make sure to like make sure to comment like and subscribe to DataFramed and make sure to check out our content for data literacy month now onto today's episode. Andy, it's great to have you back on the show.

Andy Cotgreave: Fantastic to be here at Adel. Thanks for having me on.

Adel Nehme: Awesome. So I am excited for you to be joining us for data literacy month to talk about data visualization, creating effective dashboards, how it connects with data literacy, and much, much more, but before, maybe for the folks who haven't listened to our first episode together, do you wanna share a bit about your background?

Andy Cotgreave: So my name's Andy Cotgreave, I'm Senior Data Evangelist at Salesforce having been at Tableau since September 2011. And I'm the co-author of The Big Book of Dashboards. And I also host a video series called ChartChat with my co-authors and Amanda Makulec, who runs the database society. And in suite in ChartChat we nerd out every month about the ins and outs of charts we see in the news and the cross-media. Yeah, so I've been in this field for over 15 years and just really am passionate about data.

Why are Data Visualizations Important?

Adel Nehme: That's really great. And I highly recommend for the audience check out ChartChat. It's very fun. Now, before deep diving with you today on data visualization and dashboards and storytelling, I wanna set the stage for today's chat by really trying to contextualize data visualization within the broader data literacy conversation, data visualization is often called the gateway drug to more complex data skills and tasks. So can you walk us through, maybe in your own words, why data visualization is so important and why learning these skills is also really important?

Andy Cotgreave: Yeah, absolutely. So I'm gonna attempt to explain why seeing something is important through an audio platform. So it's a bit of a challenge, but go with me listener and go with me again. Imagine you're looking at a spreadsheet of numbers, right? I don't know, sales of products across different regions, loads of numbers. If you are looking at that spreadsheet, can you see, which is the highest selling region and product can you see, which is an outlier? Can you see, which is the worst performing well? Maybe you can, but on a spreadsheet of numbers, maybe there are a hundred, or 200 digits on that spreadsheet. It's gonna take you minutes to do it. And you're probably gonna do it. Inaccurately the power of data visualization is taking aggregated versions of those numbers, or even just highlighting numbers in a table so that you can see the information you want to see in milliseconds or less. And that is what we're trying to do with data visualization, take spreadsheets of numbers or databases full of digits and express them visually. In a way that answers questions as quickly as possible. And yeah, we're gonna get deep much deeper into that

Adel Nehme: And harping on the importance of that skill or that concept of being able to understand data very quickly. In some sense, this has never been more important today. Thinking back about the past two years, for example, if you wanna look at a table of COVID-19 data spread and how it's evolving, that's not gonna be, that's gonna be tantamount to very horrible public safety messaging. So being able to showcase that with the chart has never been more important.

Andy Cotgreave: That's so true with, do you know, you think we've lived through two, nearly three or getting on for we're approaching our third year of COVID and just think about all those government press conferences. We sat through all those charts that the media and the medical professions put out in a way of communicating really complicated data in a way that educates and informs a nervous population. Right. This was fundamental. To the pandemic and hopefully inspires all to, I think, better about data literacy.

How Design Thinking helps in Data Storytelling

Adel Nehme: So now that we understand the importance of data visualization, why is such an important skill to master as part of a data literacy journey, whether an organization or an individual? I think it's really great now to talk about really what makes it effective data visualization. I'd love to dig into the details with you in your book. Provides a lot of great inspiration for that. So in your book, the great book of dashboards you lay out really well, the foundational elements of effective data visualizations. You're also someone who really borrows from the world of design thinking the world of design visual imagery to improve data visualizations and communicating with data. So maybe to start first, start off, can you discuss the different ways design thinking helps?

Andy Cotgreave: Yeah, absolutely. I think my biggest inspiration here was a book called the design of everyday things by Don Norman seminal book on engineering design and user experience that really applies to data visualization, too. Something I hadn't appreciated when I first got into this field, the big summary that Dunman says is designers make pleasurable experiences. And you might think what, how does that apply to a. Well, imagine a boring bar chart bar charts. Aren't boring obviously, but imagine a boring bar chart. If you can.

How do I bring design theory into that in order to make the user get the maximum? Outta that bar chart. Well, I can add the correct title. I can add a title that asks or answers the question that the bar chart reveals. I can ensure that the data points that I want people to see are highlighted in such a way that they see the longest bar or one particular bar.

I can do that by softening the way the, axis are formatted, making them a bit like gray or something. So you can use annotation layers and that's just on a single bar chart. Now advance that into a maybe more complex chart, you might put on social media or dashboards. You communicate with your organization there. You've gotta create this pleasurable experience in some way that matches the medium, be it social media or your business intelligence server internally. And it's still got to get the right information to the audience in the shortest amount of time. Possib. And again, Don Norman talks about how we process any designed object, whether it's a remote control, a Kele or a chart, an dashboard, we look at it, and we have a visceral response. We make a judgment based on its appearance. We have a behavioral response. Can this chart actually answer the question? We came to it with. And then we will reflect, um, did it look good? Could I answer your question? If the answer is yes, then you did a good job, and bringing design theories, laws, or theories and rules from the world of design has really taught me how to get those three levels of processing right in sharps and dashboards that we can build.

Adel Nehme: And in some sense, nailing those levels sort of processing really enables you as someone who's creating a data visualization to get action from your stakeholders, from your audience, because otherwise, if you don't nail that, the objective of your data visualization is necessarily gonna be achieved. It convinces a stakeholder, enabling an action.

Andy Cotgreave: Yeah and the keyword though, is you the objective of the visualization, you know, a common. I see all the time is, oh, they want to see sales data. Okay, well just press, press the button. I've got a pie chart of sales data, you know, what do they want to know about sales data? What are you trying to communicate? And if you haven't thought about the objective of the visualization, then it won't be a successful design.

What Makes a Great Visualization?

Adel Nehme: So expanding on the notions that you initially laid out here in the book and the different thought leadership you create as well through blog posts through your channel, your podcast appearances, you always talk about the different elements of an effective data visualization. So can you walk us through maybe in more detail, what makes a great data visualization? Great. And what are the different elements of such a data visualization? 

Andy Cotgreave: My true answer? It depends..uh, so that's generally my response. How should I visualize this chart? It depends. Okay. Now that's a little facetious or perhaps not very helpful for people just starting out on their journey. So if you'll allow me more than two words, I'll carry on. I'll definitely do that. Okay, good. So I think in order to measure the success of visualization, you've got to be asking, it goes back to what we just mentioned. What are you trying to achieve? And I think. Something I'm honing in on is a model that, that there are four tensions or levers that you're trying to pull and push and pull, push, and pull whenever you build a visualization.

And when you think about those four tensions, you are then able, and whether you've implemented them successfully, you are able to judge whether visualization is great. So the first three would be. Is your objective to show a large amount of detail or just provide the gist of the information? The second one would be, is this chart gonna be for fun or is it serious?

Right. You know, that's a valid conversation and the third tension would be, are you trying to show something where people can explore the data? Or are you trying to explain a story you've already found? The fourth tension is, is your visualization formatted, honestly or is it formatted in a deceptive way?

So that last one, we always need to keep our formatting, honest people do create deceptive charts sometimes deliberately, sometimes accidentally, but we do have to, obviously, we should all be leaning towards honesty, but the other thing about those tensions is really important. So imagine. I'm doing a presentation to the board. And I know a slide is only gonna be on screen for about two seconds and I'm gonna show it and then move on. In that case, you have very little time to convey very little information. So a successful chart, there is super simple with a, like an in-your-face message. Contrast that with a business user, who's got time to explore the data on a business dashboard that you've. That dashboard could have 10 different charts on it. And some of them could have loads of data points and there could be interactive actions that create a story and a flow. In that case, it could be really complicated. It's obviously gonna be quite serious and it's very much an exploratory experience.

So both those scenarios are generating success. But the way to know if you failed is if you take that complicated dashboard and put it in a PowerPoint. For two seconds and say, as you can see the dashboard shows sales are going up and then you move on it's Nope, I can't. So that great dashboard could be appalling if used in the wrong place. So that's why it depends because you've got all these little levers you have to push and pull and eventually you can judge success based on where you were trying to achieve.

Adel Nehme: Yeah, and I love that you used the word lever because the way I imagine this when you're breaking it down is that you have kind of this panel with different knobs, that you can evaluate the different tensions. And depending on the inputs that you have, right, is this an audience that only has five minutes to listen to your data story or a presentation? What's the medium by which you're sharing it determines the level of where you need to put the.

Andy Cotgreave: Yeah, absolutely. And I'll give you an example. Hans Rosling. He was Swedish physician who exploded onto the Ted talk scene back in 2006, showing this amazing chart of the health of nations. Right. And basically, his Ted talk was him talking about a scatter pot. And I do this exercise in presentations. I put on a slide, with 150 dots on it, and lots of different colors. And I asked my audience, is this too complicated for a presentation? People are like, yeah, it. It is. And then we play the Hans Rosling video where he presents the same chart, this hundred and 50 dots multiple color chart with two measures, animations, and it's mind-blowing the difference being hands Rosling takes the time to explain what each axis means.

He focuses on one dot. He tells you what one dot means, and then he explains what the context is within the greater picture. And then he narrates the chart sort of animates through time. So. What he's done, there is going, does my audience need the detail or the gist? And what he's realized is he wants to push the lever so that they do get detail.

The audience does get detailed, cause it's a complicated chart, but he realizes in order to achieve showing the detail in the presentation, he has to commit 3, 4, 5, and 6 minutes to what the chart is and tell people what they're seeing. So that I think is an example of how. Somebody can use these levers to achieve something pretty powerful

Adel Nehme: And maybe giving another example here, flipping the levers. Can you think of an example where the lever is more on the gist side of things and it's a bit serious? Or can you give us maybe another example where it's more just that?

Andy Cotgreave: So another of my favorite examples was a chart originally made back in 2012 by Simon Scott. And this was a chart showing conflict-related deaths in Iraq from 2003 to 2011. Right? So not a happy data set right now. I'm gonna try and explain this. The chart that Simon published was a simple bar. Bars were pointing down and they peaked in the center of the bar and then he colored them a deep blood red, the appearance, as you're looking at this bar chart with a sort of inverted triangle, looked like a smear of blood dripping down the screen, and the title was Iraq's bloody toll. So what Scott did was use orientation, pillar, and title to create this visceral response, to think deeply about the human tragedy of what happened in Iraq, in that period. Now, what I realized you could do is if you flip the bars the other way up, change it blue, you actually begin to see them, number of deaths per month, over five months is decreasing.

So in that case, you could actually change the title and say, deaths are on the decline and try to tell a story of hope instead of focus on the tragedy. And this is where the lever is really applying on your design. You're using those design levers to actually change the message in a story completely just with color and orientation.

Impact of Pre-attentive Attributes on Data

Adel Nehme: That's an amazing example and definitely, we've rushed in earlier examples where you showed it to us, especially on a webinar that you attended with DataCamp. And I highly recommend the audience to check it out. There's one section in your book that I love, which covers something called pre-attentive attributes and data visualization.

And it touches upon a lot of the notions that you discussed here. I think these provide a great framework to think about how data visualization is perceived and how to best construct one. Can you walk us through maybe what are pre-attentive attributes and how they impact a data visually? Impact on the audience.

Andy Cotgreave: Yeah. I fell in love with data visualization back many, many years ago. And it was part of learning about what cognitive science that really turned me onto it. And basically, our millions of years of evil should have hunted gathering and trying to spot tigers, moving through the grass of the Savanna and avoiding the red, dangerous poisonous berries enabled our visual system to, we could process the natural environment around us. Before we consciously think about it. And that is a gift of evolution. Wow. So avoiding tigers and finding berries enables us to be better data analysts. Yes, it's true. Because think about a bar chart, a bar chart has rectangles that are different lengths. So length is an example of a pre-attribute. And so now what actually is it that our brain looks at the different lengths of those bars?

And actually identifies, which is the longest and which is the shortest before we even look at the bar chart consciously. So we've already got a head start to the data before we actually think, what am I actually looking at or area you can make circle charts, a big circle. We could pre attentively see as bigger than a smaller circle or colors or hues.

So if you're looking for a red dot amongst the gray dots, you're gonna see the red dots pre attentively. These preattentive attributes are the atoms. From which we build chart, length, position, color, hue, size angle, there's loads of them. And it's probably beyond the scope of this podcast now, but we process some of them better than others, which is why pie charts.

Aren't very good. Cuz we don't really do angles and areas very accurately, but our brain can super accurately see differences in length of bars, for example. So yeah, preattentive attributes. Once you understand those, it unlocks so much.

Adel Nehme: Yeah, I couldn't agree more and you can actually leverage them to your advantage while delivering a presentation. For example, to guide the audience's attention by using these pre-attributes, a great example would be, you know, if you wanna point the audience's attention to one bar in your bar chart, you can elect to make everything gray and just highlight color on that bar chart, mid presentation. 

Andy Cotgreave: It's so easy to misuse color in visualization or any communication medium because every tool available to us today can use an infinite amount of colors, but the most powerful visualizations dashboards are the ones that use gray and one color. And then really, really powerful. 

Adel Nehme: Okay.That's awesome. So of course, given the book is called the big book of dashboards. I'd love to actually deep dive with you on dashboarding. As dashboards are one of the most effective ways to share insights with data visualizations within any organization today, you know, many organizations are leveraging tools like Tablo, click power BI to do these dashboards. So can you walk us through maybe how dashboards extend the power of data visualization and what you have found are the best practices for creating an effective dashboard?

Andy Cotgreave: Well, I think first we have, we have a semantic challenge. In fact, I'll ask you Adel how would you define a Dashboard.

Adel Nehme: So I would define a dashboard as a collection of data visualizations aimed at answering a specific set of questions around a specific set of data within an organization. How is that definition?

Andy Cotgreave: It's great. Right. Have you thought deeply about that in the past, or is that, is that your first stab?

Adel Nehme: That's kind of my, my first stab at it, but I'm, I'm not a data visualization expert as you Andy.

Andy Cotgreave: Oh, I'm sure you are. I'm sure you are. Okay. So in the Dashboard. Can be defined in a gazillion different ways in our book, the big book of dashboards, our definitions is only 15 words long, and it's super vague because what we realized is, as we were looking at all the dashboards that we could find across industries is that there were so many different variations. So one example, you said a collection of charts signal your example, but we've got some great dashboards that we think are dashboards. They're just a single visualization. So every time we try to extend our dashboard definition, We could just find more and more caveats to be like, oh, well, this dashboard doesn't fit the definitions.

All we're gonna do. So in the end, we just collapse the definition to something pretty vague about it being an artifact. You used to monitor a system and facilitate understanding. I think it was something like that. The reason I don't really know what our definition was is that I don't really care about the semantics, right?

You or the audience. We are trying to collect information and present it to a user in such a way that they can make decisions. Check a process or understand more about whatever it is they're looking at. Right. And if you wanna call that a dashboard, great. Ultimately a dashboard is a word referring to a piece of wood on a stagecoach.

Anyway, so it's a word taken from somewhere else. So now we've got a sort of definition. Well, now I've maybe destroyed the definition of the dashboards. Who knows? I dunno. How to best create one. Well, you have to go to your audience and really understand very deeply what it is. They want to see if they say, oh, I'd like to monitor sales.

Oh, great. Why would you like to monitor sales? I mean, it might be, are we on target is our quote on target this quarter? Why do you wanna measure that? And they'll come up with a different answer and believe me when you go to users and if you ask why four or five times, you know, this is classic of business, MBA process, you'll get to the root cause of what they want once, you know what they want. You're like, well, how do you want to see it? Are you gonna be interacting with this thing? Do you want this thing delivered in an email? Are you gonna be looking at it on a cell phone or on a screen in a call center? Each of those will determine a different delivery mechanism in a different style. So the summary of this, the best practice for developing effective dashboards is go and speak to the user and understand what they want, why they want it and how they want it, and then create a really basic prototype. And then they'll go, no, that's not what I want at all. And then from there, you can iterate until you get to the right answer of what they want.

Adel Nehme: That's really great. It's definitely complicated, but it's also wonderfully simple and accessible, which is what's so nice about dashboarding and data visualization in general.

So the book contains a lot of examples of dashboards from different industries and different use cases. And you showcase brilliantly why these dashboards are effective. Can you walk us maybe through the different type of dashboards that you've encountered and maybe expanding onto that? What makes each of those dashboards effective?

Andy Cotgreave: Yeah, I think a big question you have to ask is should the dashboard be interactive or not? And I I'll focus on that for this answer. So if, if something's gonna be interactive, then you've gotta start asking, well, do my users understand how to use this dashboard? How do I make sure they literally know how to use this platform?

You know, for Tableau server, Tableau cloud, for example, they need to know what the URL is. And then once they get there, what is it they're looking at? Even when you put filters on a dashboard? How do the users even see them right now? This sounds so ridiculous. Adel that I could say, well, it's on the screen.

Surely they'll see them. Well, we've done a bunch of eye tracking studies on dashboards. You know, a lot of which would taken from the big book dashboards. And I could have designed a dashboard with filters on it. Then we allow, allow people to look at the dashboard themselves and they literally don't even look at the filters.

So then at the end of the exercise, you ask them, well, why didn't you interact with this stuff? And they go, I didn't see the filter. And the inside I might be screening, but they're on the screen. , they're literally in front of you, but understanding that people look at screens in a way that you might not predict or might not be as you'd hope.

So you've gotta ensure they can see the things to interact with in the first place. So interactive or not is important. We have a great example in the book from arsenal football club, one of the premier league teams in England, and they have this static. Based on a player's performance, which is delivered to the player after each match, but just before the training session and it's a static dashboard delivered to their cell phones so that they could show that to their teammates and have a laugh or have some serious insight, but basically analyze their own performance.

And they're really fundamentally different types of dashboards, cuz one's interactive. This is an example that isn't. The, the really lead to the question is like, well, is this dashboard effective? It goes back to the previous question. Are we thinking about how the user gets it and what we're trying to achieve?

So, yeah, I think interactivity or static is a big decision to make. Another one is what is the form of delivery? Is it gonna be primarily used on a big screen? Is it gonna be primarily used on the cell phone or is it just gonna get delivered in an email again that they will require completely different form factors that you have to take into when you're designing your D.

Considerations for Dashboard Creation

Adel Nehme: That's really great. And maybe deep diving here a bit more. What strikes me from your answer is there are really two main considerations people need to have when thinking about the dashboards they need to create, one is the audience, right? What is the audience expected to achieve with this? And then secondly, what is the format and the user experience of consumption of dashboard in an organization?

What are the different types of audiences in a nutshell, basically of personas that may expect to interact with a dashboard, someone is developing and what is often the ideal way of presenting that inform. 

Andy Cotgreave: The types of audiences, it's difficult to try and summarize that as a catchall. But one example I often see is your executives, right?

Let's think about let's. I mentioned sales earlier. Let's think about sales, an executive CEO or head of revenue. They want to see what is the aggregated roll revenue this quarter, compared to last quarter, compared to target and compared to this time last year, right? They wanna see this thing rolled. Up to a very high level of aggregation and they'll have KPIs, which will show whether they're on target.

And then, uh, you know, some slightly disaggregated breakdown, maybe by region or by product area, something like that. That dashboard is useless for the account executive who is actually trying to use data to target an account. So an account executive completely different experience. You know, maybe they have five or six accounts and one of the accounts is a leading car retailer or something, right.

The executives dashboard is. Utterly useless to this person. They have to be able to take the same dataset and say, well, maybe they can use data to tailor an account plan to target that account. You know, what have the users in that organization been doing? Have they been looking at our website? Have they been doing some training courses?

What have we sold to them previously? Or are there any opportunities that have been one or five recently? And so it's the same data set, but the dashboard, the account executive needs is completely different to the dashboard. The exec. And the reason I use this example is sometimes we see in organizations, the executives go, we have done business intelligence.

We are successful because I have an executive dashboard. Hey team, everybody use the executive dashboard because it reveals a complete misunderstanding of data, culture, and data literacy to think that, Hey, the dashboard you've got is great. What about all the people believe in the organization? You gotta think for those people too.

So that's one example.

Adel Nehme: That's really great. I'm excited to expand on the data literacy component here, and maybe on the user experience before we move on to the next question, what are things that you wouldn't expect need to be considered as part of a dashboard design process? That would be from a user experience perspective.

So for example, one thing that came to mind, I was reading an article about this recently is just how important low times are. For a dashboard or to be able to be consumed. Right. And this is not an interactivity decision, not an aesthetic decision. You could have one of the most well-designed dashboards of all time, but if it takes more than five to send seven seconds to load, it could really hurt the number of times it's actually used.

Andy Cotgreave: Yeah. And I think we might have questioned later about if database is the gateway drug, what else do you need to consider? And this is where these things have to be considered. I know in, in Tableau, for example, if you build a Tableau dashboard and keep it bare bones, super simple, and you've got a well formatted data set and a big, nicely resourced server, the low time will be really fast.

If you get carried away as a designer and start thinking, I'm gonna bring in loads of bells and whistles and background images and do all really bespoke calculations.

Adel Nehme: And I had a lot of interactivity, for example.

Andy Cotgreave: Yeah. Yeah, yeah. Then you're actually then beginning to put more of a burden on the server and that dreaded low time begins to increase. And this is one of the challenges, certainly, something, when I used to be an analyst, when I was a customer of Tableau before 2011, I used to get carried away with building elaborate dashboards that are really intricate, but they took forever to load. So you have to sometimes. Recognize that the enthusiastic designer that is inside you trying to build these wonderful experiences has to be balanced with the need to create something that actually doesn't lose people when they're trying to load.

It's a great question. Important thing to think about. Yeah,

Embedding a Narrative in a Dashboard

Adel Nehme: that's awesome. So of course, the other side of things here beyond user experience and beyond design is the ability to create a narrative, right? Communicating data insights and data storytelling. It's extremely important when crafting data visualizations and dashboards. So can you walk us through maybe how to effectively embed a narrative within a dashboard and how to convey that insight to a consumer?

Andy Cotgreave: Gosh, I've been resurrecting a talk. I did back in 2014 about a dashboard I built, which was very much inspired by. DBER right. What DBER now think about DBER comic strip. The weekday strip is a three-panel comic strip panel. One is an introduction to the dope joke. Panel two is the joke. And the third panel is like some sort of epilogue that hopefully adds and builds on the joke itself. That's an amazing story structure. Introduce the plot, complete the plot, write an Epilogue right.

I built dashboards. That follows that story structure, three-panel dashboard oriented horizontally, introduce the data, deliver the punchline and create an apple log right now. That was for a particular bespoke situation. But the reason I'm using this example is because dashboards are made of panels containing charts that are in some sort of a grid-like structure and they should be in some sort of a grid-like structure. And where do we see that? We see that in comic books. Right. And what do we do in comic books in the Western world? We read them from the top left to the right. Well, we read them from the top left to the bottom, right.

In a, sort of a linear structure. So if you want to form a narrative in a dashboard, a really good starting point. Is two thing like a comic strip. So the top left again, I'm talking about Western left to right reading coaches here, the top left contains the super summary. And then from there you can lead the user left to right.

And down into more detail. So whatever's in the bottom, right. Is kind of the most granular level detail, which is where you've got the deepest level of context. Now that's not a universal way of designing dashboards or way of doing things, but that's one example, go and read some comics, basically.

Data Literacy through Visualizations

Adel Nehme: I highly recommend as well. There are incredible mediums to convey storytelling visuals. So of course now we covered kind of the main skills when it comes to creating effective dashboards, visualizations, and narrative. But I think given that it's data literacy month, I also be remiss not to talk about the connection between data literacy and data visualization, right.

From an individual perspective. And from an organizational perspective, you know, I said at the beginning of our episode together, as you alluded to as well earlier, data visualization is often called the gateway drug to more complex data tasks. Right. So can you walk us through, maybe from your perspective. How consuming data visualizations and learning about data visualization enables better data literacy within organizations and more thoughtful conversations around data. 

Andy Cotgreave: Well, I think when you have good data set and your building charts, you know, your dashboards for example, are successful when users come to you and ask more questions, but, oh, I've seen the sales by region.

Thank you very much. Now what's happening by product, but that's a sign the day. A gone and looked at your visuals, understood it. And then it's inspired a second level question, and that is a sign of success. You can't always answer every single second level question. You shouldn't aim to do that because you can't anticipate all the questions users are gonna have, but that's a sign.

People are engaging, right? So that's on dashboards. As you start bringing data into presentations, or even just into meetings where you're throwing data around on screens in an ad hoc. You can start querying data very, very quickly and getting answers instantly. I hear customers a lot. Say one of the problems we have Andy is we spend 15 minutes of E every meeting arguing about the data, and I've always thought isn't that kind of the goal.

And I realized it's a problem. If the argument about the data is cuz you don't agree. With where the data's come from. And you don't agree about the truth of the data. Obviously we've gotta solve that, but I'd love to meetings to be, let's go in.

We're gonna try and continue sales. How are we gonna improve sales this quarter? Right. Well, let's play with the data. Let's explore which bits are underperforming or outperforming, and let's have that conversation and argue about what the date was saying to come out with the decision. So I think that's a really high level of maturity where data is just the fabric of the conversation.

And data is easy to access easily understood by everybody and driving questions as they arise, which dashboards can't do. Dashboards can only answer questions you already thought about what are you gonna do to answer? Today's. Yeah,

Adel Nehme: That's really great data, visualization skills. What's so nice about them as well, is that they give you that confidence to be able to criticize a data artifact, right?

Because especially visualizations, they tend to present themselves as a matter of fact, as ground truth, right? Because it's beautifully visualized it's there in your face. And having those skill set really equips you to become much more constructive and also much more critical of data.

Andy Cotgreave: Yeah. That's, that's so good. And a really important part of that is empathy as well. I'll tell you a little story. I used to run a blog called deciphered reality.com and it analyzed data about the board game, ARCOM horror, ARCOM horror, the card game, right? It's a nerdy collectible card game where you build decks to take your characters through a scenario against Lovecraft inspired monsters.

Hey, one awesome guy, right? There's an entire website dedicated to which cards are used in which kind of. And I started building charts based on the data in this website for fun. Right. And because it was, I love the game anyway. So I built the charts. Here's the most common used card. Here's the least common used card, blah, blah, blah. And then somebody on RA replied. Yes. I don't think you understand data visualization analytics should be about this, that and the other. And you are only sharing the first level of insight I recommend you do. Thanks for your advice. I'll think about that when I write my second book and when I appear on another set of podcast and when I do another set of keynotes and teach another 10,000 people, but what that person had failed to do.

Show the empathy of, well, what was I trying to achieve? I was just trying to achieve a little bit of fun, looking at super basic chart. I'm going, here's the basic thing. Take the data for what it's worth and have some fun. And he had kind of gone well, he's seen that. And then he is gone. Well, I've gone.

These are the 10 questions. Dammit. He's not answered these questions. He's a failure. I only answered the first question. Now you go and do the work if you are so bothered about it. So anyway, I can now after three years of hurt, laugh about that story, but the point being. That person was criticizing in a non-empathetic way.

And I think data lit mature data literacy knows how to criticize, thinking about the design's intent and which what's the actual goal of the visualization they're critiquing. Right. So, oh, there you go. Shared my story.

Adel Nehme: yeah, that's really great. I appreciate the vulnerability. So given how important data visualization skills are, you know, what are the ways you recommend to people within an organization to become better at visualizing data and also consuming data visualization?

Andy Cotgreave: Well, I guess, first of all, I'd say listen to the DataCamp podcast, get involved in data literacy month and all the resources that are there for you, right? So that's a given, but also just practice, practice, practice, practice, data visualization, and getting good at building dashboards and being able to communicate effectively with data.

It is an art and a skill it's technical, it's art related. And you know, an artist is not successful from day one. An author is not successful from day one. A Coda is not successful from day one. They have tried, they've played, they've succeeded, they've failed. And every single step they're taking is teaching them a little bit more.

There are free tools such as Tableau public. You could just download it or use it on the. Connect to data and get going. And every time you build something, you are learning on top of that, there's really rich and active communities. You can get involved with one example in the Tableau community is something called back toss basics.

Every two weeks. It's like here's a really basic data set and the challenges build a bar chart or build a scatter pot. It doesn't get easier than that. And so the barrier of entry is super, super low. Cuz you can learn that skill in about half an. But then you can connect to all the people who've done it the same task that month.

And believe me build a scatter plot with this data set will generate infinite different types of scatter plots. Right? So practice get involved in free tools, obviously Tablo public isn't the only one. And then essentially steal like an artist still like an artist was a great book by Austin. Cleon oh my gosh, I've forgotten his name, but his manifesto was in order to become a good artist.

Or in this case, a database designer, you can go and get inspiration from expert practitioners and sort of copy their work in a way which is not plagiaristic right. Copy something for inspiration and with humility. So I think get involved, download something for free and. Be inspired by the working others.

Exciting Future Trends

Adel Nehme: That's definitely the case. And I really appreciate these insights and advice. And now Andy, as we near the end of our episode today, I'd love to ask you, what are you up to next? And where can audiences find you and given your position at dub low, what are future trends and releases in the business intelligence space that you are excited about?

Andy Cotgreave: So I've recently got promoted. So I'm senior data evangelist at Salesforce now, and I'm excited to bring data visualization and data culture, understanding to Salesforce, customers, and prospects, which is a really big platform. So I'm excited about that at time of recording, I'm gonna be going on a sabbatical soon.

So at the end of that, I'm beginning to work on a new book. Can't say very much now, but you've heard me talking for an hour and you've got an idea about what I'm passionate about. So just keep your eyes peeled on that. So that's me. And then future business intelligence trends. I think what we are seeing in our area is that Tableau and other tools like power and click create tools.

Which in the hands of expert analysts can turn data into anything. What I think Tableau hoped 10 years ago is that anybody can learn to use Tableau. What we've learned is that actually analysts like to use Tableau or people who are analytically inclined like to use Tableau, right? So they put in the effort to learn the platform.

However, not everybody has that inclination to spend the time learning the ins and answers using this platform. So it's how do you. Power of say Tableau or any analytics platform to people who, for whatever reason, can't invest the time into learning that drag and drop experience. And we're doing it through things like ask data, which is a natural language interface to Tableau.

And we've been working on it for years and the latest iterations. I are even beginning to tempt me, you know, a 15 year veteran of using Tableau away from the native Tableau interface to, I'm just gonna type a sentence like mine and Google, and then tinker with the words in that sentence to tinker with the view. So I think our goal is to try and bring the power of analytics to people who aren't analysts, you know, that's things through our data and bringing data to the user, rather than asking them to go to a destination to see. So, yeah, that's the trend I'm excited about.

Adel Nehme: It's very exciting to see the trend of conversational interfaces and all sorts of analytics tools.

Like even you see it in open source, programming languages like Python, you now have something like KOD deck and you say, Hey, I need this data set. I need a function that creates that. And it creates it for you as well. So this is gonna be very exciting towards the democratization of data and the democratization of data insights

Andy Cotgreave: I D have you been using the AI generating artwork? Yeah,

Adel Nehme: I have. I have DHI too. Yeah,

Andy Cotgreave: yeah, yeah. Sorry. It's dally, not ley. I keep calling it Wooley, which is, yeah, they're amazing. Right. And that kind of AI, that language interpretation, you know, that's what we're trying to bring to data as well.

Adel Nehme: That's awesome. So, Andy, it was amazing having you back on the show. Is there any final call to action before we wrap up? 

Andy Cotgreave: I guess I tell first of all, DataCamp data literacy month. I think what you're doing is fantastic at DataCamp. So I support everything that's going on there. Second, just start your own journey. If you are thinking about starting your journey, just download a dataset, try it and see what you can find from it. And I guess if people want to follow me, I have a newsletter called the sweet spot, which will get renamed soon, but. Slightly to do with the book, but currently call the sweet spot. And I'm sure we'll put links to that in show notes.

Adel Nehme: Awesome. Thank you so much, Andy, for coming

Andy Cotgreave: My absolute pleasure. Thanks Adele. I think you're doing a great job and I hope you all enjoy the data literacy.

You've been listening to data, framed a podcast by DataCamp. Keep connected with us by subscribing to the show in your favorite podcast player. Please give us a rating, leave a comment and share episodes you love that helps us keep delivering insights into all things. Data. Thanks for listening until next time.