How To Analyze Data Through the Power of Data Visualizations

Natalie Takacs:
Good morning, everybody. Welcome to how to analyze data through the power of data visualizations. My name is Natalie Takacs I’m the Product Manager for the Connected Analytics products Report and Explore at InEight. Today, I’m joined by Catie Williams, Director of Connected Analytics at InEight. And Mike Moore data viz expert and data scientists at Aurora Cooperative Elevator Company. We’ll get started with some intro. So, Catie I’ll kick it over to you if you want to get started.

Catie Williams:
Sure. So, as Natalie said, my name is Catie Williams. I have been in the reporting and analytics space for about 10 years now. I am the Product Director at InEight and we are responsible for creating solutions for construction and engineering software. If you’re not familiar with InEight, that’s what InEight does. And prior to that had experience at Kiewit. And then I also am the program director at Bellevue University for both their graduate and undergraduate data science programs. So, I love all things, data and database.

Natalie Takacs:
Excellent. Mike.

Mike Moore:
Hi, all. My name is Mike Moore. I’ve been working in analytics and reporting for over 15 years, almost 20 years, but I’ve been really focused on data visualization and the study of that for the last 10 years. And I currently work at Aurora Elevator Company where I’m a data scientist. Prior to that, I worked at Kiewit Corporation where I got experience in the construction industry. And then, my early background in analytics and data visualization was in communications and working with voice recognition systems and analyzing responses to those types of software.

Natalie Takacs:
Awesome. Thanks. So, you both obviously have a background with data and analytics. What specifically drew you to being interested in data visualization? And what would you say is your favorite part about that?

Catie Williams:
Well, let Mike go first. He’s got more experience.

Mike Moore:
Yeah. So, I think the thing that got me interested in it was really two things. So, when I was in school, I was always very interested in art and creative things. And I think that the data visualization aspect of it, the creativity and really the ability to build a picture from data appealed to me and I was kind of drawn to it just because it was a creative outlet for me, that was the first thing. And then, the second thing was really some of the results that you could get from it, a shared understanding of the messages that you’re trying to communicate. Like I’d mentioned, I started my career in analytics and measuring call center data and then voice recognition data from automated IVR where people would call in and speak to the machine and we would analyze that data and then try to make sense of it and try to tune the applications so that callers would have a better experience.

Mike Moore:
And so much of that work was done with Excel spreadsheets and with access databases. And then, when we have conclusions or results, we would try to put those into PowerPoints. And that was a very laborious process. It took a lot of effort to do. And when modern data visualization tools like Tableau and Power BI came available, and the tool that I was first exposed to was Tableau. It made that process much simpler and it also was very easy to create and bring data in and create different visualizations. You could fail fast, you could decide what worked and what didn’t work, and you could do that all very quickly. And so, just the speed of development and the ability to control the narrative and create a shared understanding of what the data is saying was what really appealed to me about it, and it was exciting. It was fun. And like I said, it really gave me a creative outlet.

Catie Williams:
Do want me to go, Natalie?

Natalie Takacs:
Yeah, Catie.

Catie Williams:
I mean, I think everything Mike said is accurate. My path was a lot different. I was asked to implement a brand new tool with a big ERP implementation. And so, that brought a lot of visual tools with it, but I wouldn’t really say that it wasn’t really, until I started at InEight, that I felt personally really drawn to the power and the impact that visualizations could have, because I traditionally had been asked just to produce that Excel-based type of reporting. I mean, that was very, very common. It was what people said they wanted, but then I think, as part of us being able to build our own product and provide customers the solution that should help them and help their analysis, that’s when it really kicked off for me.

Catie Williams:
I was fortunate enough, several years ago, probably like 2011, 2012 to go to a Stephen Few training. I would call it a training. It was more than that, more than a conference and seeing like the passion around it, different philosophies around color and being minimal, but really purposeful was very impactful, but it wasn’t until a while later from a group perspective that I really felt a lot of passion personally around it. And with all these tools, just like Mike said there so much has become possible so quickly, that barrier to entry has really been reduced where anyone can do this. And so, I think that part of it being, I wouldn’t call it easy because there are some challenges, but it’s really easy to get started. And then it’s really fun.

Catie Williams:
And just like Mike said, I think the excitement you feel when you show something someone, and they can’t believe that you were able to do it. And that part has really been fulfilling to continue to increase my interest in this space.

Natalie Takacs:
It’s really rewarding to see that aha moment when someone can see an insight that you were able to bring to them easily through data visualization. One quick thing I want to mention to the audience a housekeeping item is that you do have the ability to ask questions at any point in time through the question box in the go-to webinar. So, we’ll be monitoring that question box throughout. And so, if you have any questions, feel free to put them in at any time. Let’s get into the meat of this. So, what is your favorite historical visualization, Mike? And what impact do you feel that that has had on the field of data visualization today?

Mike Moore:
Sure. Well, I think the first thing I’d say is that there’s been a lot of talk and a lot of content created around data visualization and a lot of that has happened in the last 10 years. It’s modern data visualization tools, have become more prevalent and easier to use. And I think that gives the perception that data visualization is a new space. It’s like new, but I kind of separated. Data visualization is actually very old. And what I call on-demand data visualization with these modern tools. That’s newer, that’s what’s new, but the data visualization has been around for a long time. And I think the first one I’d want to talk about if you could bring it up is the… There’s a book that the Menard System.

Mike Moore:
And if you go to the next slide, this shows a graph that Charles Menard made about the Russian, or about the French march to Russia and then the retreat. And basically it shows, Charles Menard was an engineer in France. And he did a lot of this work in his retirement. This is the most well-known visualization that he does, but when this book came out, it highlighted a lot of his other work that’s pretty impressive. So, this visualization shows that the thickness of the line, of the tan line is the size of the army of Napoleon’s army, as they’re marching out to Russia. And the black line on the way back is the size of the army as they’re coming home.

Mike Moore:
And you can see they got smaller and smaller through the brutal winter on the march to Russia and the defeat, and then the retreat home. And there was just a small fraction of what’s left. And there’s two things to take away from this. This is number one, visually gives you a path and an idea of the size of their army as they went out and on the way back. But it also, if you start thinking about the human costs of what this actually says, and these are not just lines on a chart. These are very impactful. If you start thinking about these, these are actually lives that are decreasing as you go out and then exponentially sends you go back and it can be really powerful if you understand what it is.

Mike Moore:
And it’s a lot more powerful than just putting numbers on a graph like, “Okay, today we had this many yesterday and we had that many.” It’s difficult to cognitively see where the losses are occurring and what the magnitude of those losses are. So, I really like this chart for that. And then if we go to the next one, this is another bit of Menard’s work. And what this is showing is cotton imports to Europe, at three points during the mid-1800s. So, 1858, prior to the Civil War, the blue mark is cotton exports from America to Europe and England specifically. And then, there is a little kind of line down to Europe there. And then if you keep falling the blue, and then the tan comes from India, I believe. And you can see prior to the war, England gets almost all their cotton from the Americans.

Mike Moore:
1864, 1865, the Civil War is going on. There’s a blockade of Southern ports. The South is busy doing other things, they’re not exploring. And you can see how that trade dwindled down but look where they picked up the slack. Again, it comes from East Asia. You can see that there. And so, it’s just very interesting to visually see now, if you need precision from that, you probably want to look down at the table and numbers, but most of the time, this will show you what you need to know is that before the Civil War cotton trade from the Americas was very heavy. And then during the Civil War, once the blockade went around, it was down to a triple, pretty much.

Mike Moore:
And so, I thought that was very interesting. So, that’s one of the historical business that I really liked. And there’s a couple of others that are worth mentioning. I don’t know, Catie, if you wanted to jump in here and mention these or not.

Catie Williams:
I mean, I think, when we were talking about this before the webinar, you and I both felt like Menard, I know I personally feel very interested in Menard because I’m in the construction engineering industry and he was an engineer. And I think that’s pretty neat that someone with that civil engineering background has so much context of data and statistics and visualizing how well those play together. And I think I see that a lot in our industry that engineers make great analysts of data or presenters of information. And then, he also talks about being heavily influenced by W.E.B Du Bois. He’s got several visualizations here that I’ll just quickly jump through, these are all hand done. And it’s really neat to be able to see how someone… Just like Mike said, these visualizations have been being done for a very long time, and now it’s just a lot easier to make them, but very impactful visualizations that were manually done, that have existed for a really long time.

Catie Williams:
And then, I think we have Florence Nightingale too right here. So, another example, and I don’t know, Mike, can you specifically wanted to call anything out about this, but-

Mike Moore:
I would just say, oh, sorry. I would just say that this would be like a tough visualization to make with any tool you would probably need to use some type of coding or scripted language to do it or something like Photoshop not Photoshop but-

Catie Williams:
Illustrator or something.

Mike Moore:
Illustrator, because you’d still require a lot of custom work. But what this does, and there’s two things I’ll mention about this. The first one is that so this is diagramming the causes of mortality during the Korean war. And so, what Florence Nightingale was trying to do here is she’s trying to not only show the different types and categorizing them and high chart light visualization, but she’s also trying to give a little bit of a time component to it, or give some time series analysis and show you which months these occurred and the quantities. And the second thing to mention about this is, is that not only are these hand-created, gathering the data was often done manually by hand.

Mike Moore:
And there’s another visualization, we don’t have an image for it, but people might’ve heard of it. It’s the John Snow visualization, where he’s analyzing the cholera outbreak in London, and he created a little map and then, he put dots on where the wells were, and he’s one of the leaders in modern epidemiology, and he’s got dots where the wells occurred and then he’s got marks on the households of where deaths occurred from cholera. And the most deaths occurred closest to the wells. And they determined that previously, they thought it was bad air, and it turns out it’s bad water. But he went door-to-door and gathered that data. And so, it was not only a manual process to create it, it was also a manual process to gather the data.

Mike Moore:
And now that we have modern technology, we have sophisticated databases and data collection mechanisms. And then, we have tools that are easy to connect to the data and basically draw meaningful pictures of the data quickly. I think that’s the reason why in the last 10 years, you’ve really seen this explosive growth of data visualization as a field of study, or a discipline, something that you can build a career path around.

Catie Williams:
That’s great, Mike. And I’m going to stop sharing.

Natalie Takacs:
One of the things that I think made those early visualizations seem so sophisticated, to me at least, was the use of color. Like those maps at the beginning that you showed, they should one color for the path out and one color for the path back. And that helps you differentiate what you’re looking at. Are there any philosophies you have on using color or any best practices that you like to follow?

Mike Moore:
Yeah. So, Catie mentioned Stephen Few, and Stephen Few book Information Dashboard Design, I’ll show it right here. Right there. I’m not plugging them, but I am a little bit, I guess. So, he’s got a number of books. That was the first book I read on data visualization and really, it’s about designing good dashboards and he’s got a lot of do’s and don’ts. And so, what I found is that what I think are best practices for me are things that I try to follow, tend to start there. And then, everything I’ve read after that has built on that. And his work is not necessarily unique originally. He built on the works of other folks like Edward Tufte and William Cleveland, other people who had written on the discipline of data visualization, he was just the latest voice at the time, but he was very thorough and he had a very subtle, clear do’s and don’ts for visualization.

Mike Moore:
So, I started with that. And when it comes to the topic of color, his advice was use color sparingly, and avoid the overuse of color. And so, what he means by that is, “Don’t use color just for the sake of using color. When you use color, you should have some meaning behind it.” Now, you can take an extreme view of that, where all my visualization should be monochromatic. And before I had read Stephen Few, I just thought when I see these really big, colorful dashboards, I think, “Oh, that’s really cool. They have dials and different things, and a lot of color on them.”

Mike Moore:
And I thought, “Oh, that’s really cool. It’s neat.” And then I started reading and started thinking about what works and what doesn’t. And then, I started thinking, and then I started seeing visualizations that more adhered to that philosophy. It were color, had a specific meaning, and it made a lot more sense to me, and they became more aesthetically pleasing to me too. So, for instance, if you’re going to show, I think this is a question we’re going to talk about, but showing negative value. So, if you’re going to show positive and negative values, and what is the information that you want to pop out of that? And so, if the information that you want the user to be aware of is the negative number so you can go fix them. That’s what you want to pop out of there.

Mike Moore:
So, you may have the bar chart or a tree map that’s monochromatic, or has fairly muted colors, but then the negative number will be perhaps a red or a gold or something that stands out, that catches the eye, that draws the viewer’s eye to it. So, when you use color, you should have a purpose to it like does it tell a story? Does it communicate the information that the viewer needs to see? And so, that’s my philosophy. This doesn’t have to be monochromatic, you can use multiple colors. You want to limit the number of colors you use, but make sure that they have a purpose and they don’t overwhelm the viewer with color, because it’s like, if you have one or two colors, that means something, they mean something, the viewer’s eyes are drawn to it. But if you overwhelm the viewer with color, then too much color doesn’t mean anything.

Catie Williams:
I mean, just to add onto that, I know that I’ve and I went to Stephen Few and I think that is a great base to start from. I have typically gotten feedback if I use all grace that I think there’s almost like an immediate knee-jerk reaction to not think that it has what you need. The information isn’t available or what you would want someone to get across from it isn’t clearly seen, although I don’t know that I agreed with that thought that that’s actually how it is, but I do like the shades of gray personally. And I know that I try to think about accessibility and I think Natalie, you’re going to ask us about that. But I think, recognizing that a large population is colorblind so red and green is going to be a really common colorblind color where someone’s not going to recognize that.

Catie Williams:
I like the blue and orange, that’s common in some tools for the positive and negative. For us that goes with our product colors. And so, we take a little bit of advantage of the marketing capability that we can get by having our visualizations styled around our product suite. Another I think really important thing that I think you should consider is how we read in the US, where we’re reading left to right. And so, that realize that you’re taking has you place things and where people’s eye gravitates towards, is really important as well, in addition to color. But just like Mike said, I think really taking a step back and asking yourself, “What am I trying to get across? What is the first thing I want someone to see when they look at this visualization?” Is really important.

Catie Williams:
And I would agree that if it’s got a ton of color, it’s distracting and it’s very, very difficult to know what is it actually telling me? And this is my issue with pie charts, and I could go down a whole path with that, but it’s very hard to figure out which piece of the pie is bigger than another if you’ve got a lot of them. And so, you have a lot of colors, I think it just becomes overwhelming and difficult to understand your message. And every visualization should be purposeful, in my opinion, it should be the right visual to display the right information. And so, color plays a big piece in that, for sure.

Natalie Takacs:
So, it sounds like the goal of color is to direct the user to where they should be looking or what they need to know, but also placement of visualizations and types of visualizations. There’s a lot that goes into making a dashboard or report really functional and usable for users so that they can get the most out of the data. So, it sounds like we’ve covered accessibility. Are there any comments either of you have on maybe colors or any other pieces of a dashboard for accessibility?

Catie Williams:
I will add data labels, I think is another really quick win to provide more accessibility. So, clearly labeling things and I think that you could have way too dense of labels, so that should also be done in the right way, but I think providing labels and tool tips, things like that are a good option. I have seen some pattern adding for the colorblind sometimes, you might hash mark something in a visual. I don’t think that’s as common, but I definitely think data labels is a big thing to help. And I think Mike, you were going to say something.

Mike Moore:
Yeah. I was just going to say that, if you’re doing a scatterplot visualization where you maybe have different points and then a few different categories, and then you’re going to use a different color for each grouping, perhaps you could also use a different shape, like use a circle for one, a triangle form another or an X for another. That would be one thing, or if you’re doing like a KPI dashboard that you may have a user requirement that I want my KPIs that are positive, I want them green. And I want the ones that are negative, I want them red. And you may have a user who’s insistent upon that or an owner or somebody you’re doing the work for, but one way to sort of… If you can’t get around the red green, one thing you can do is use either an up or down arrow in color that red or green.

Mike Moore:
So, even if the viewer has a red, green deficiency, they can still look at the shape of that mark, and then tell what that means. And then, a lot of software tools now, some will have the ability to put a graduated mark in there or line in there or bar but some of them don’t. So, if you want to make sure that you have some ability to be accessible for people with color deficiencies, you can use different values as well. So, like make sure that even, even if two pillars are similar for somebody who’s color deficient, if they are different intensities, you can tell them apart.

Mike Moore:
And there are tools out there that you can put on your screen to tell, how would it look to somebody who has a color deficiency, especially red, green, which is the most common. It’s really important to make sure that you do that because, I don’t know the exact statistics, but what I’ve read in coding a lot is that, about 8% to 9% of the male population that has a red, green color deficiency and about 1% to 2% of females still. So, if you’re using red and green, you’re basically making it very difficult for 10% of your audience to understand what you’re communicating them to. And I remember I did a lot of work interviewing people who have had this red, green color deficiency.

Mike Moore:
And one person I’d asked, I said, “How did you know that you couldn’t tell the difference?” And he goes, “Well, I was hunting with my dad and he said, “Dad, where’s the deer? I mean, I can’t see where the deer went.” And he goes, “Where you’re standing.” And there it was and he pulled the blood, he couldn’t tell, it looked green, simple as grass and he couldn’t tell. That’s when he knew he had a problem. I had another friend like that, who all the lights on the traffic lights, they just look like white light. He couldn’t tell. And so, those are examples of people who do have trouble navigating in a society where green is up and red is down or positive or go or stop, those are challenges.

Mike Moore:
And if you can address those accessibility issues in your dashboard, you really make sure that 10% of your audience that would have been excluded before can consume the information you’re trying to present.

Catie Williams:
And the other thing we didn’t mention, but I think also consistently using color the same way. So, if you choose to do, like in my case, the blue and orange for positive and negative. Well, then I shouldn’t flip it, and then I shouldn’t be using it for something else. So, I think that consistency, so that if you do have individuals that have maybe found some workaround can have this consistent experience when they’re consuming your information, and you’re not constantly flipping and flopping where this means this status, but now on a different screen, it means a different status, for example. So, I do think when you’re using color, that’s something else to really keep in mind is, should be purposeful and consistent so that you don’t introduce unnecessary confusion.

Natalie Takacs:
Sure, yeah. Mike, you talked a little bit about using grays and red to indicate positive and negative trends or indicators. Catie, you discussed using orange and blue. Do you have any other tips or suggestions for indicating positive and negative facts and figures?

Catie Williams:
We use the up and down arrow a lot, and Mike you said that I like it. I like it better than the plus and minus but consistently I’ve been using that up and down arrow.

Mike Moore:
The other thing I would mention is, you need to be aware of the type of visualization that you’re doing. For instance, if you’ve got negative values, a pie chart, it’s not really going to work very well for you, because you can’t really display those negative numbers, unless you separate the pie out and put it over there and slice over here, but it’s a little difficult to do so. So, you want to make sure, you’re using the right visualization or if you’re using a line graph or a scatterplot, if you’ve got negative numbers, you want to be able to clearly mark where zero is and, so it’s important the type of visualization that you’re using or that you’re doing if you’ve got a lot of negative numbers in there.

Natalie Takacs:
And it probably goes back to what Catie was saying about utilizing data labels too, being able to indicate this is a negative number, maybe not necessarily utilizing color.

Mike Moore:
And that’s one area that doesn’t get talked about enough is the texts that you include with your visualization. I mean, you can have a bar chart out there, but if you don’t have meaningful, understandable texts, you can lose a lot of the context behind it. So, the data labels can be very important, especially with positive and negative and things like that.

Natalie Takacs:
Definitely. So, that dovetails into my next question. So, that’s a definite do of data viz. Do you have any other do’s and don’ts, that are pretty critical when creating data visualizations?

Mike Moore:
Yeah. I would say, for me the first thing is what are you trying to communicate? So, if you’re trying to communicate an amount, and we’ll take the pie chart example again, it’s very difficult for viewers to tell proportions with area. If you have a slice of a pie and two slices are kind of similar in size, but one’s a little bit bigger. So, it’s kind of hard to tell how much bigger that one slice is than the other. We’re much better when we view data or view visualizations. We’re much better at discerning differences from link than we are with area. So, when you’re trying to communicate amounts, bar charts are more effective than maybe a pie chart, for instance.

Mike Moore:
Another thing is that I try to avoid is using too much precision because it can slow analysis down. And then we’ve talked about the overuse of color. Catie, did you have a few?

Catie Williams:
Yeah, I mean, you mentioned that axi. I think that’s important, you should always start at zero, but if you ever watched news, if you’re ever reading like an article in the newspaper or something or online, you’ll notice how often they don’t start at zero to try to sway you, because you’re just looking at it, you see this huge difference, but really if you put it at zero, you would see that there’s not a significant difference. Sorting seems obvious. You’d be surprised how often people don’t sort either ascending or descending, but it makes a huge difference in your ability to quickly see what’s the greatest and least. If you don’t add sorting, you are putting that on the user to have to go and do that. We talked about labels, which I think are really important, the data type.

Catie Williams:
And we’ll talk about this a little bit more later, but there’s better ways to show time, than showing categorical information. And so, I do think there’s some visuals that work better than others when whatever type of information you’re trying to display. So, just like Mike said, understanding what you’re trying to convey is extremely important. I would go so far as to say a pie chart is a don’t most of the time, unless you have a Boolean, so a true false. The only time I think it really works well, I do think it’s nice to have some visual diversity if you’re building a dashboard, it is nice to not always just have 10-column charts and they’re all the same, but I also think that it can be more distracting than just adding that visual difference, but if it is a true false, I think the pie chart or donut chart is what the eyes can prefer, can work, but it definitely isn’t my go-to by any means.

Catie Williams:
We’re going to talk, I think, about running totals, but I’m a big fan of having that secondary axi, if you could have a running total line, I know that there’s a lot of schools of thought that says that’s a big no-no, because it’s too confusing for the user, but I am a big fan of seeing my actual values and bars in a cumulative in a line. And now, I’m going into specific visualization. So, I’ll stop. Sorry, Natalie. It’s easy to go to-

Natalie Takacs:
Oh, that’s all good. I think that transitions nicely into my next question. What are your favorite visualizations?

Mike Moore:
That’s for Catie.

Catie Williams:
screen share?

Mike Moore:
Yeah.

Catie Williams:
Let me share my screen. All right. Let’s see. All right.

Mike Moore:
So, this is what’s called a bullet chart and it was designed by Stephen Few. And you can Google it, and there’s a Wikipedia page on this. It’s also on his website for perceptualedge.com, but what it does is it basically looks at where you are now versus where you were or where you want to be. And then what percent, are you of achieving that goal? So, in this case, the black bar is the value that you’re measuring, the horizontal bar. And then the vertical black mark there, is a symbol marker that is the comparative measure. So, whether you’re trying to compare to the prior year or prior year to date or if you have a goal set, that can be the goal. And then, there’ll be gray, different shades of gray backgrounds in this case are different percentages of attainment.

Mike Moore:
So, if I think about the Tableau visualization tool, when you create a bar chart, it’ll let you set percentages. It will be like 60%, 80%, 90%, or a 100%. And then, you can color this. Now, what I’ve been doing a lot is I will get rid of these gray backgrounds because I think it’s a little bit cluttery when you have that too much and I’ll just have the bar and then the vertical reference sign of where I’m trying to get to. And a lot of times what I’ll do is I’ll do that and then at the end of the bar, I’ll put what the current value is. And then where that vertical line is, I’ll put what is the goal value that we’re going after.

Mike Moore:
And then, I’ll get rid of the scale at the bottom. And so, makes it a little bit less cluttered and puts all of the information, the text right next to the visual. So, it gives you the context that you need, but it it’s a good way. This is a really a good way to show, “Here’s where I am or here’s where I’m going and here’s where I need to get to, or what I need to exceed.” It’s a good way to compare values.

Catie Williams:
Before you go to your next one, I think if you ever want to look at your annual results and look quarterly, that gradient in the background, I think, helps provide a really quick way to see are you meeting each of your quarter goals? But I would agree, Mike, that I think the bullet chart does so much really well, but sometimes it gets really busy. So, you do have to explain it, but then it feels like once you’ve explained it, people love it and get it. And I know that you and I both have been asked to expand it, to add additional data points to it, to make it even more powerful but it displays a lot of information for sure.

Mike Moore:
That’s the thing about data visualization. Sometimes too much information can be too much information you want to try to keep the message clear and not clutter it and lose it with unnecessary stuff. So, this next one, we talked about a cumulative line chart. So, this is something that I’ve been doing a lot lately and I’ve been combining it sometimes with the bullet charts. So, in this case, I’m looking at a set of numbers. This year is the darker blue, last year is the prior year, the lighter blue. And then I’m looking at values by week number. So, week one, week two, week three. And so, if you could see where this lighter blue is, you could say, “Okay, last year I seem to have more sales earlier in the year.”

Mike Moore:
Let’s say this is sales. I had more sales earlier in the year and I work in agriculture. So, things are very seasonal and sometimes depending on the weather and the conditions, crops will sometimes get planted later or farmers may apply fertilizer later in the year, some years and other years. And so, this will show you that. So, if we’re looking at something here and I don’t think there’d be any product, but let’s say we’re comparing this year to last year that we don’t really start reaching a peak sales until week 28 or 29 here. And we started gradually attending that earlier in the prior year. So, what’s going on here? Why is there this gap between these two lines or from week 30 on down, and then did that hurt us overall?

Mike Moore:
Well, in this case, you can kind of say, “Yeah, it did.” The prior year we overachieved or we sold more than we did this current year. And we flattened out earlier too. So, there’s a lot of information there that you can research and then you can come in and start looking at, “Well, what were the conditions that caused this?” And it can give you a place to start looking and researching, why something is working or not working. And so, one thing that I’ve been doing, I mentioned I’ve been using this together with the bullet chart is, I will do a visualization like this. And then, in a small multiple format where I have several of these for different metrics that I’m trying to measure.

Mike Moore:
And then above it, I’ll put the bullet train and that just shows the total. So, you don’t have the week by week time series aspect of it, but the total will give you the quick view of where you are right now, and then the view underneath. And then, I’ll put it underneath and I’ll put this line chart and it’ll show how you got there. And so, if you add seasonality to it, it can show that very easily. And the other reason I like this running total, as opposed to just doing a line chart and just showing the individual values for that week, or that month, that you’re measuring is a lot of times you get charts where these numbers go up and down, and it doesn’t really tell you a story so much. If you need to measure month to month differences, a bar chart might be a better way to do that because you’re measuring quantities from one month versus the next. But if you’re looking at a trend over time, I really like this cumulative trend.

Mike Moore:
And then, if you have some negativity come into the, if you start losing something and then you’ll start dropping down and you can really see that. If you have a cumulative chart and it starts to go down then I’ll have a couple that do that, then you know you’ve got a real problem somewhere and you can see quickly, and you can communicate that very quickly as well.

Catie Williams:
That’s great. So, one I selected is the tree map. I think in our industry, in the construction engineering industry, we are interacting with hierarchies a lot. And I think a tree map is a really, really good way to display a lot of information in a hierarchy. And I have another example where I have a bar chart here or column chart or whatever you want to call it that would typically be used. So, I’m showing like productivity data by a category of work, essentially. And I think the reason why a tree map works a lot better in this case is, I don’t know if you noticed that the largest one right here is the building, which is an 83. And if I come over here to the tree map, I don’t even see that code, that 83 is nowhere here.

Catie Williams:
And that’s because it’s in the bottom right hand corner, one of those bright green boxes. So, it’s not really a lot of work. So, how this visualization works right now is like the scale of the work. So, the number of hours, for example, in that code is the size of the box. And then the color indicates how it’s doing or not. And so, it would be really easy to be misled by this chart to thinking like, “Oh, building’s doing so great.” But the truth is it’s doing great, but it’s not a significant amount of the work. So, should I really spend a lot of time and do I really care? When really I might have a bigger problem in my mechanical or my electrical and instrumentation, and then where this gets even more powerful is this little white line, which I think is hard to see.

Catie Williams:
And I don’t think you can see my mouse, but there’s a white line that separates each of the boxes and that’s the percent complete of the code. So, I can tell how large the scope of work is. I can tell how it’s doing from an overall productivity perspective. And then, I can also see how far along it is. So, that tells me, if it’s really close to being complete and it was red, it’s unlikely I can do a whole lot about that to fix that operation or to make it better. Whereas on this code 81, and it might be hard to see the white line on the screen, but it’s showing me that I’m a little over half of the work, so I could actually start to impact that code and hopefully start to change my productivity.

Catie Williams:
And then when I mentioned that it’s good at hierarchy information. So, this works in a hierarchy, I’ve got mechanical equipment, and then there’s going to be a bunch of codes below that and more below that, but I could then drill into one of these codes. And so, I’m going to drill into 81 because it was not as far along, and it was a little bit more red, but I can now drill into that code and then start to analyze, “Okay, well, of that overall category, where can I now focus more attention?” And so, this tree map gives you so much information and it did take me a step to explain it, but now I can quickly, if I’m looking across projects or even at an overall level, I can see where do I have the most issues across my organization? Where do I have the most issues on a project?

Catie Williams:
And then quickly get all the way down to the specific operation to hopefully make an improvement. And I did break my red, green rule here, but this was specifically asked to use red, green. And I don’t think it’s terrible. But I mean, again, maybe using gray and red would be a good alternative or gray and blue, for example. But I love the tree map and I’m talking a lot. So, Mike, I don’t know if you wanted to say anything before I go to the next.

Mike Moore:
What did you just say about that last example where you… So, this first one here, so you have the red and green there, that’s going to probably be difficult to tell, but if you go down to the next one, for some reason. This might not be so difficult because you got different intensities here. So, you want to think about that. And that’s where tools that can help see where you see what somebody who has a color deficiency can see. Then you can adjust your values that way. So, it’s not always wrong to use red and green.

Catie Williams:
[crosstalk 00:42:07] with the blue and orange on this one, but then the middle of that is a brown, which is an interesting visual.

Mike Moore:
Yeah, that’s why I struggled with blue and orange. You know what, I use a lot is I use blue and red a lot, or I’ll use blue and different colors, black and gray. I tend to use those more than just because the dark colors of orange are not very good.

Catie Williams:
So, then the other one I wanted to call out and I don’t think this one is as popular, but I really like it again, once you’ve been explained this chart, I think it’s really easy to quickly spot things that are related. So, this is a chord chart. I think it has other names as well, but if you look at the big bands, that’s where you want to focus your attention first. And so, over on my left-hand side, I have my type of incident. So, I’m looking at something by safety and then over on the right are all my different industries. And so, I can see the type of incident first aid, I have that the most of any of my incidents when I compare it to a non-occupational or a no treatment, lost time. And then, I can see from an industry, I’ve got the most incidents occurring in industrial.

Catie Williams:
So, if you follow those chords, you can see, where am I having the most first aids? Oh, it’s in industrial. And then you could see vertical building is next. And again, the bulk of their incidents are going to be first aid, but you can follow those chords. And so, when you look at like this one, now it’s really easy to quickly see, oh, transportation has the most change orders. And so, I like from a visual perspective, how fast it is to see two things that are related to each other, because you don’t really care about the precision like Mike was saying, but instead I’m focused just really quick like, “Okay, where should I focus while change orders.” And then RFIs would be the next place or health and safety because those bands are larger as well. And those were the only two I had, Natalie.

Natalie Takacs:
So, this brings up an interesting question. You both showed some really impactful visualizations, but both mentioned that it does sometimes require a little bit of explanation to understand immediately what you’re looking at. Do you think that’s a barrier to people utilizing these more impactful visualizations? Catie, I remember when we were looking at the chord chart for the first time, once we figured it out, it was this amazing aha moment. And we thought of so many different ways we could utilize it. Do you think the fact that it requires some explanation holds people off to using it?

Catie Williams:
Well, I was going to list it as one of my don’ts. We were going through our do’s and don’ts that if it requires explanation, it should probably not be used, but I think that it just depends. I mean, every time we’ve shown this chord chart, people immediately grasp it, they get it, they understand it the next time, same for the bullet chart, same for the tree map. So, I think the benefit outweighs the little bit of explanation. And I do think through annotation and other tools, you can actually start to overcome that barrier. We don’t primarily use annotations a whole lot, but I think that’s a way to get past that. And I don’t know, Mike, if you have other ideas or if you think the ones I shown should not be used.

Mike Moore:
No, I think, one of the things about the chord charts that you mentioned is we haven’t talked much about interactivity, but with the modern tools, there’s a lot of interactivity there. So, if you want to focus on one aspect of that chord chart, a lot of the tools that you would build that with Power BI, for example, you can click on that mark and then it’ll highlight the peaks that you want to see and then put everything to the background. But as far as instruction goes and having to learn, so one of the emerging fields of study and there’s companies coming up around it is around this concept of data literacy is the word that’s being used a lot. There’s actually a company called Data Literacy where it’s a newer company founded by a person who came from Tableau.

Mike Moore:
Ben Jones is his name, and good guy, smart guy has been part of the Tableau public community for a long time, led that group for Tableau. Now, he’s gone off on his own. He’s created this company called Data Literacy, and he’s got a couple of books out there. It’s got a new one coming on, but the whole idea is that as a society, we are not very good at reading charts and graphs. You get that when somebody says, “Hey, can you build me a pie chart for this?” And you think, “Okay, well, what are you trying to say with it?” And then go, “All right, I’ve got 15 different categories here and I’m supposed to squeeze those over and do a pie chart with 15 different colors. And they’re almost impossible to read.”

Mike Moore:
And so, what are some of the other options? So, now, you’ve got things like line charts and bar charts that we’ve talked about, and those are very basic, but still require a little bit of instruction and you go all the way to something more complex, like the chord diagram that Catie showed that does require some explanation, but just because it takes a little bit of learning to do, and maybe it takes a little bit of study where you have to look at it where, I used to have this idea that a good data visualization should be something that you can look at it and instantly know what it means. That’s not necessarily true, a good data visualization needs to be clear and fairly intuitive, but if it’s a complex subject, it may require a complex visualization.

Mike Moore:
But it’s the whole literacy aspect of it. Do people know how to read the visualizations that we’re creating? So, it’s not bad to have to put a little bit of education behind it. And I think Catie, you mentioned consistency with colors, like when you’re using negative colors, they’re always orange or something. I think consistency around the visualizations that you create is what will help promote the literacy within your organization. So, one of the things that we’ve tried to do and in my organization is we’ve tried to make sure that we’re always continually using the bullet chart concept and this running total time series. Here’s where we are and how we got here. And we use that a lot and in that literacy, the more we use it, the literacy of that, of our organization when reading those trends continues to improve, because they’re familiar with it, they’ve seen it before. And it’s just a little bit of education, but it’s not bad to have to do that.

Natalie Takacs:
Sure. So, how do you think data and data visualizations have changed the way people make decisions?

Catie Williams:
Well, I mean, I think there’s a lot more than just the visuals that go into it, but with technology it’s made information available much faster. So, I now no longer have to go and pull data out manually, put it into a spreadsheet and aggregate it and then create a chart. So, I think just from a speed of making decisions, visualizations helps tremendously, but so does technology. I mean, but I also think when Mike was talking about that cumulative chart, if you have that in a table and you’re having to like go line by line, it would be so easy to miss something or to, I mean, just from making an error, but it also would be so time-consuming to have to go through and try to figure out, “Well, should I go up, should I go down?”

Catie Williams:
So, I just think that speed of, there was an answer really quick, but a lot of it just goes back to what Mike was saying about the literacy of data. And so, if you don’t know how to read it, then data visualizations might not help a whole lot, but overall I think that they have tremendously increased your ability to make a decision, much more timely.

Mike Moore:
And just one thing I would add to that is that, Catie mentioned spreadsheets, if you give somebody a table of data, you give 10 people the table of data spreadsheet to look at, and then you come back to them. And you give them a day to look at it, come back the next day and say, “Okay, what did you get from that?” You’re going to get 10 different answers because people are going to look at it and draw their own conclusions from it. But if you’re an analyst and you’ve got a finding and you want to communicate that finding, you’re not going to throw a table of data at somebody, or you want to make a change in the organization and you need to make a decision, you can’t have 10 different people coming to different conclusions.

Mike Moore:
You need to create a shared understanding. And that’s the one thing that I think data visualization can do that. I don’t know if it gets talked about enough, but it can help drive a shared narrative or shared understanding. Like, “Here’s the problem. And we can see this visually and how do we address that? And so, I think that shared understanding of what the information is saying, that you can use visualization to promote that shared understanding. Now, conversely, you’ve got to be careful how you do that. There was one of the authors that I read a lot is Alberto Cairo, who’s a data visualization professor at the University of Miami and his latest book is called How to Lie in a Chart or something like that. And it’s all about how you can mislead with charts.

Mike Moore:
And it’s basically a book on the best practices, the do’s and don’ts of data visualization with kind of a catchy title but you could see that Catie mentioned, using a zero baseline for bar charts, are you propping them at some place to maybe overemphasize the difference that maybe really isn’t there. So, you got to be careful about what you’re doing and stick to some of the best practices so to speak. But I think data visualization can help create that shared understanding of what the information or what the data is telling us.

Natalie Takacs:
You both mentioned utilizing more tabular style, pixel perfect reporting like Excel. Why do you think people are still utilizing this kind of reporting?

Catie Williams:
It’s lack of trust, I think is the primary driver and the need to touch it themselves to see the details. I mean, I think it all just comes down to the trust in the system, trust in how it’s doing the math, how the data is coming out and just having that focus of control.

Mike Moore:
I would agree with that. I think, it comes down to comfort. I mean, we’ve talked about, we showed examples of how data visualization is very old, but on-demand data visualization is very new. And in the last 10 years, it’s become very prevalent in a lot of organizations, and I think most of it, I mean, those of us who are past 45, grew up in our roles as analyzing spreadsheets and looking at tables of data. And now, a lot of people who reach their mid 40s or above are in leadership roles. And they’re comfortable with tables of data and they don’t necessarily trust the visualization as Catie mentioned. And then, they’ve got to see those numbers behind it. They’ve got to touch it.

Mike Moore:
And then, I think the other aspect is a question of precision. I think people tend to sometimes want to be more precise than they need to be. Now, let’s take an example, like let’s say sales are down by $5,000. Is it enough to know that? What do you need to know that they were down by $4,943.25? What do you need to know? If you’re an accountant, you probably need to know that precision. If you’re the CFO, you probably need to know that precision because you’re going to have to sign off on the financials at the end of the year. You need to know that, but if you’re in operations or if you’re in sales, isn’t enough to just another, “Hey, I’m down 25%. I’ve got to do something. Or I’m down $5,000.”

Mike Moore:
How much precision do you meet to make a decision? And I think people think they need more decision or more precision than they do to make a decision. And I think that’s part of it as well.

Natalie Takacs:
And that makes sense. So, how would you suggest to someone to drive the change, to use data and adopt data visualization across their organization?

Catie Williams:
I think, what we have been trying to do is provide it both until we can get the business process to change and the comfort with what we’re producing to remove the tabular. But that’s typically what we’ve done is like, we’ll give you some of the detail to make you feel comfortable while also providing you hopefully a route that will make things faster. And then over time, it seems we can start to remove that dependency on the tabular.

Mike Moore:
And then, there’s a concept that… Catie you’ve heard this phrase, the art of the possible and it’s like when you try to explain data visualization to somebody, it’s really hard to communicate what you mean, but when you show them, you get a much better response, when you’re showing them exactly what you mean. So, it’s like showing them the art of what is possible to do, what can you do? And then, I think the other part of it is building champions and grow. Every organization where data visualization has been adopted and promoted, it’s grown organically. In a previous organization, one director starts doing something, they go to their directors meetings’, you get a call from another says, “Hey, my director was with the director from down here and he saw this and now he wants to do this or she wants to do that.”

Mike Moore:
And it’s like, “Okay.” You’re starting to get some momentum from being able to do things, show information in a way that is meaningful. And once people start seeing it and are able to understand it and use it, then you can really start going to some adoption with it.

Catie Williams:
That’s great.

Natalie Takacs:
And along those lines, what would you recommend to somebody trying to get into the field of data visualization? Are there skillsets that are beneficial to have? Or do you have any resources that either of you could share that are particularly useful for somebody looking to grow their data visualization skillset?

Mike Moore:
So, I would say curiosity is probably one of the best skills to have, just as a desire to be able to learn and then be able to communicate what you’ve learned to others and do that in a visual manner. Knowledge of basic statistics helps and sometimes that scares people who are not inclined to that. And you’re like, “I came more from an archie side of the equation.” But there’s data visualization has people from many, many different backgrounds, some are very mathematically-driven or scientifically-oriented. Others are creatively-oriented and it’s all helped to enrich the space. So, basic statistics, you can learn that having some basic understanding of that and understanding regression and things like that are helpful. Also, just being able to tell a story and understanding what storytelling is because while we don’t necessarily want to steer somebody in any direction, that’s not true, we do want to use some type of storytelling points to try to illuminate what is actually happening.

Mike Moore:
And then, how’s it represented by the data that you’re using? I think as far as how to get better, if you’re new to the field, practice, there’s a lot of different tools and resources out there you can practice and it’s just practice, build a portfolio and share your work. I think we’ll talk about resources in a second, but just do it and you can get data sets from different sites, like Kaggle or data.world. And there’s all kinds of free data out there and you can go out and visualize with, and practice and then share your work with other people. Build your portfolio.

Catie Williams:
I mean, Mike took all the words out of my mouth, so I don’t have anything. And I know we’re almost out of time. So, do you want me to share, Natalie really quick our resources.

Natalie Takacs:
Sure, yeah.

Catie Williams:
So, we just put together a few, if you wanted to take a screenshot of it, just some of the resources and Mike talks about the community that’s there and available. All of these different resources from Visualizing Data, Perceptual Edge are great for you to find additional information. And there’s a whole community of people that can help weigh in. And if you build a portfolio, get feedback. So, these are some of the ones that we follow and would recommend.

Natalie Takacs:
Awesome. Well, thank you, Catie and Mike, thanks everybody for joining us today. Feel free to reach out with any questions. Have a great Wednesday.

Catie Williams:
And if anybody had anything that they’d like us to talk about in another session, feel free to let us know. We’d love some feedback.

Natalie Takacs:
Awesome. Well, thanks so much. Have a great day.

Catie Williams:
Thanks, Natalie. Thanks, Mike.

Mike Moore:
Thank you. Bye.