Get tips and short practical techniques on topics such as infographic design, forecasting, and t-testing.
- [Voiceover] Welcome to Data Science Tips, a weekly series that is hands-on demonstration and guidance for those working in data science. I'm Nate Makdad, a full-time analyst and data advocate and will walk you through topics that have been very valuable to me. We hope you like what you see and invite you to share with others that might be interested. Each week we'll cover a new topic and provide references to other courses in the LinkedIn Lynda.com Data Science Library that might be of further interest. Let's get started. Hi there, this is Nate Makdad and I wanted to talk today about applying a visual test to charts.
A visual test is a way for us to take a step back and think about how easy or difficult it is to view and interpret the information that we're presenting. The reason this often comes up is because, have you ever seen a hard-to-understand chart or report? Maybe too much information or color that got in the way of each other? This is a pie chart that I found just by googling around for charts, and there's a lot of competing information here, really around the colors. The colors can get a bit distracting, and so, though, even though it can hard to distinguish the size between the bars, there's also just a lot going on here, and this is a chart that would take a lot of time to really digest and understand.
And so we want to speed that up and visual tests can help us with that. So visual tests are really about making sure the layout and elements work together, that it can keep our layout from getting crowded or messy, and that, as we take our step back, we might decide that there are certain elements that need to be modified and/or taken out because they're visually distracting and making it harder for people to see and understand the data. Really, this is always about making things as easy for the viewer as possible, so that people can see and interpret the data as fast as they can.
So the first test I want to talk to you about is where are your eyes drawn? And this is really a pretty simple test, but is intended to give you a chance to step back and see where a user will be drawn on the page. It's really trying to take advantage of some pre-attentive processing principles, known as Gestalt principles, and we really want to be paying attention to colors, shapes, and layout, seeing how colors and shapes may draw our attention or how we could redesign layout to improve the look of the chart or dashboard.
And this is really where we want to make sure we use this design to help guide our viewers through the process. So to conduct this test, really you just close your eyes for three to five seconds, or just look away from the screen and then we want to just be thinking about, what's the first thing you see on the page? Where are your eyes drawn? And let's go ahead and use an example. This is from a blog post called, Where Are Your Eyes Drawn posted on tableau.com, and I've copied and pasted the charts into here so we're just going to advance through these.
So here comes chart number one. Just to give you some background, this is a representative chart trying to compare, say your business performance against some competitors on a variety of different scales. And now, where are your eyes drawn on the page? Well, for me, there's a lot of competing information, just visually. The bright colors and the smaller dots make me look all over the page. I'm not really sure where, exactly, to start. I'm just going to start left to right, and then I'm trying to compare the business, so that blue diamond against these other ones, and so I constantly am trying to reference that diamond against the other values, and so I'm going up and down a lot.
So there's a lot of visually distracting information, in my opinion here, on this chart. And now, let's take a look at the same idea, but maybe in a different layout. So here's chart number two, and so this is trying to, again, use some of those Gestalt principles to think about how we could change the layout. And so, in this view, where are your eyes drawn? And so, for me, the first thing I'm really noticing, right, is the blue; the blue really stands out. It's not a bright color, but given the contrast against the gray, it really stands out.
It's much easier to quickly understand how the business is performing against competitors. I do lose some information about the individual competitors but, at the same time, if our primary focus is just one highlighting our own business performance, that may not be as important. Also the grouping of the colors, so because the business is such, it stands out so much better. The blues really make it very obvious, and also the single use of the ranking system really helps as well. So it's much easier to read and really navigate through it much quicker in terms of understanding how your business is performing against the competitors.
So now let's take a look at a second test called the Squint Test. And the Squint test is really a design practice to help you think about how new users will see and experience your design and layout. We're focusing here on broad shapes and colors, and we want to see, does the color, shape, and layout really help or hurt our design? Are the first things that stand out in the blur, is that where we want people to go? And really, the Squint Test is exactly what it says it is. You just want to squint your eyes together to blur the page so that you can take away all of the detailed information and just start to get a sense of the background layout and color scheme and how that's affecting your layout.
There's actually a Chrome extension that you can add as well, so if you do any web design, I post a lot of dashboards to Tableau server, so I'm often opening them up in a browser, you can actually do the Squint Test right in the Chrome browser, as well. So I'm going to use that one first, just to demonstrate the Squint Test here. So here's the Squint Test, and I've already added it to Chrome, and I'm just going to ahead and click into my chart here. So this is the same chart we showed earlier, and you can see the Squint Test here, and now I'm just going to go ahead and add a bit of blur, and really, what do you continue to see as we blur this image? Well, for me, there's still a lot of color.
There's a lot of color and a lot of directions, and so, if at the blurred level, the color's pulling me in multiple directions, then unblurred, it's going to do the same thing. So, really, this chart would probably need some clean up or some work using the blur test, because we got a lot of competing color elements. Now let's go ahead and take a look at another example, this time, an Excel-based-in example, and this is from a blog post called, The Dashboard Squint Test. And here is the dashboard that was created, and you can see there's a lot of information on here.
There's a lot of charts, a lot of numbers, and so, having spent some time on this, we may be a little concerned about where are people going to be drawn on this page? What's going to stand out? Does the layout work in terms of people being able to navigate and find the information they need to quickly? So let's go ahead and blur this. They had blurred this in the blog posting, and so we're just going to take a look at picture number two. And so, here is the blurred version of the image, and really what continues to stand out on this page? Well, first the red dots really continue to pop.
So if we're really trying to draw attention using the reds to indicate areas of concern, that definitely still works. That red stands out very clearly on the page and definitely draws attention. The white lines really help break up and divide the different section, so we can see about six different sections, and they're all broken up through straight white lines. So that certainly helps with the breakout of the different sections as well. So, really, again, when we're thinking about the Squint Test, what stands out? Which shapes and colors stand out? Do your eyes get drawn to the places you want to go to? Does it look like it has a structure and organization that we could follow? Do your eyes go where you want your viewer's eyes to go to? Or does blurring everything make it very difficult to navigate, and it just becomes a jumbled mess? And if that's the case, then maybe we need to start rethinking either our color, layout, or possibly taking out some information.
So if you're ready to learn more about visualization and design, there are a couple of courses that might be of interest to you. The first one is called Data Visualization for Data Analysts by Bill Shander and touches on the data visualization elements and fundamentals and foundations for thinking about design. I have another course, Creating Interactive Dashboards in Excel 2013, and there's a set of movies, specifically focused on best practices in dashboard design.
And then another movie that might be of interest to you is Designing with Gestalt Principles by David Hogue, and this one, we touched a little bit about the pre-attentive processing, or Gestalt principles, and so this series of movies really helps in terms of thinking about how you can use these principles to help you layout and design dashboards. So finally, applying some quick visual tests can be helpful to see and remind us how new users will experience our design, and also for us to check for unnecessary or distracting colors or images.
Tune in for interesting, relevant, and thought-provoking tips from the wide world of big data and data science.