You know what great design looks like when you see it, but how do you actually achieve it—particularly if you don’t consider yourself a designer? SWD covered four topics to help you think like a designer: affordances, aesthetics, accessibility, and acceptance. In this chapter, we’ll practice applying these concepts and illustrate how minor changes can help take your visual from acceptable to exceptional. First, let’s cover a quick reminder of what I mean by these terms.
In visual design, affordances are things we do to make it clear how to process what we show. This builds off of the lessons you’ve practiced in Chapters 3 and 4: tie related things visually together, push less important elements to the background, and bring the critical stuff forward. Direct your audience’s attention intentionally to where you want them to look.
Spending time on the aesthetics of your visuals can translate into people taking more time with your work or having the patience to overlook issues. Attention to detail comes into play: often many seemingly minor components add up to create a great or poor experience. To achieve the former, we must edit ruthlessly.
People are each different, and accessibility means recognizing this and working to create designs that are usable by people of diverse skills and abilities. We’ve touched on colorblindness, but that only scratches the surface. We’ll undertake exercises that will help you think about your designs more robustly. There is one simple thing that can help us improve the accessibility of our graphs broadly: using words wisely.
Finally, our visual designs only work if our audience accepts them and there are things we can do to make this more likely, which we’ll explore.
Let’s practice thinking like a designer!
First, we’ll review the main lessons from SWD Chapter 5.
When we communicate with data, people sometimes have the false belief that words have no place or should be kept to a minimum. But words play a critical role in making the numbers and graphs that we use to communicate data understandable to our audience. The text we put on our graphs helps people comprehend what they are seeing and can assist in shaping their perceptions about the data.
Let’s do a quick exercise to illustrate the importance of words on graphs.
Study Figure 5.1a, which shows sales over time for four brands of laundry detergent. There are already words on this graph: but are there enough? Could we use words more wisely? Consider these questions as you look at the data, then complete the following steps.
STEP 1: What questions do you have about the data shown in Figure 5.1a? List them! What assumptions would you have to make to interpret this data?
STEP 2: What words could you add to this graph to answer the questions you raised in Step 1? Freely make additions and changes to title and label so that what is being shown is perfectly clear.
STEP 3: How could putting different words on this graph change the interpretation of the data? How can you change axis titles and other text to cause an alternate understanding of what this visual shows? What implications does this have for what words should be present on every graph? Write a paragraph or two summarizing your learnings from this exercise.
STEP 4: For hands-on practice, write on Figure 5.1a or download the data or graph. Either add text to the existing graph or create a new one in the tool of your choice, practicing using words wisely to make the information accessible.
When you create a graph, the details are almost always clear to you. The challenge is that they aren’t necessarily obvious to your audience, who may have different expectations or understanding of the context. In absence of text to make the data comprehendible, your audience is left to make assumptions, just as you had to do in this exercise. Not only does this make you use more brainpower than necessary, but worse—those assumptions might be wrong!
Let me take you through my approach to this exercise to illustrate how choice of words can lead us to completely different interpretations of the data.
STEP 1: I have four main questions about this data.
Consider how different perspectives answering the questions raised above could lead us to totally different interpretations of this data. Let’s look at that more specifically next.
STEP 2: Figure 5.1b shows one way I could add words to this graph to answer the questions I raised in Step 1.
In Figure 5.1b, I assumed these represent unit sales for the four brands of laundry detergent sold at a specific store. I made this clear through titling: substituting a more descriptive graph title and adding axis titles to both the y- and x-axes.
Let’s review some specific design choices made with the text I added to this graph. I left-aligned the graph title. We’ve discussed the typical zigzagging “z” of information processing a couple of times already (in the solutions to exercises 2.1 and 3.4, as well as in SWD). As a reminder, without other visual cues, your audience will start at the top left of your graph and do zigzagging “z’s” to take in the information. By orienting our graph title at the top left, our audience hits what they are looking at before they see the actual data. This is the same reason for orienting my axis titles at the top (y-axis) and left (x-axis).
I paid close attention to detail in the alignment of my axis titles, orienting the y-axis title to align with the top of the highest y-axis label, and the x-axis title is aligned at the left with the left-most axis label. I chose all caps for my y-axis titles (and will often do this for axis titles in general). Because capitalized letters are all the same height, this creates a neat rectangular shape (compared to what you’d get with mixed case: a jagged edge). I like the framing this lends to my graph. I also wrote the axis titles in grey text, so they are there to make it clear what we are looking at, but aren’t drawing undue attention or distracting from the data.
STEP 3: Alternate words could lead to a totally different interpretation about what this data is and represents. See Figure 5.1c.
This has implications for the words that should be present on every graph. I can generalize into a couple of guidelines. Every graph should have a title. When communicating with a slide deck, I use descriptive titles for my graphs and takeaway titles for my slides (we’ll talk more about the latter in Chapter 6). That’s certainly not your only option, and we’ve looked at examples in this book where the graph title is both descriptive and highlights a takeaway. Be consistent in how you title with a given report or presentation.
Every axis should also have a title. Exceptions to this guideline are rare. Title explicitly so your audience doesn’t have to spend their brainpower trying to figure out or make assumptions about what they are viewing.
Words make our visuals comprehensible for our audience. Use them!
The graphing applications we use to visualize data are built to meet the needs of many different scenarios. This means that it’s rare that the default settings will meet the needs of any one of those scenarios exactly. That’s where we come in—our understanding of the context and design sense can improve defaults tremendously, helping make information more easily digestible and simply more pleasant at which to look and with which to spend time.
Let’s dissect a specific example, considering how we can use lessons in design to improve upon default output from a proprietary tool and create a more desirable experience for our audience. See Figure 5.2a, which shows the number of cars sold by dealership over time for a given region.
STEP 1: First, let’s simply react to this graph. What words come to mind in terms of how this graph makes you feel? Make a short list of the feelings this graph evokes.
STEP 2: What changes would you make if you needed to communicate the data from this graph? Specifically, address:
STEP 3: Download the data and graph. Remake the visual applying the changes you’ve outlined in the tool of your choice.
STEP 4: Imagine you have been asked to create a single slide focusing on this data that will fit into a broader deck to be shared with the management team who oversees these dealerships. How would that affect what you show or how you choose to show it? What additional words can you put around it to help it make sense? What other design considerations would you make? Create this slide in the tool of your choice.
STEP 1: My initial response to this graph brings to mind words like: confusing, chaotic, overwhelming, and complicated. These are reactions I’d like to avoid when I communicate with data!
STEP 2: The following describes how I would approach remaking this graph to both better get the information across and foster a more pleasant overall experience for my audience.
Use of words: I like the fact that everything is titled in the original, but I’m not a fan of the center alignment of the graph and axis titles. I would upper-left-most justify all of the titles so that when my audience starts at the top left, they encounter how to read the visual before they get to the data. I’ll choose all caps for my y-axis title because of the nice rectangular framing that this, together with the graph title, creates for my graph. On the x-axis, we probably don’t need the title of Quarter, as this is quite obvious from the individual labels. I’ll omit this. There is currently a lot of redundancy with the x-axis labels given the repeated years, so I’ll pull those out as super-category axis labels.
When it comes to creating visual hierarchy, I have to decide what to focus on in this graph. In the original, it’s difficult to focus on anything because so much is competing for our attention. I see that Regional Avg is emphasized in the original via a thicker black line (though this doesn’t stand out nearly as much as it could given all the other lines, colors, and shapes). I’m going to push everything else to the background. When it comes to eliminating distractions, I’ll also remove the grey background, borders, and gridlines. Getting rid of these non-information-bearing elements will help my data stand out more and make for a less cluttered feeling visual overall.
In terms of additional changes I would make to the overall design, it currently takes work going back and forth between the alphabetical legend at the right and the data it describes. I’d like to eliminate this work for my audience. My typical method for resolving this is to label the lines directly. This is challenging here because many of the lines are close together, but I’m still going to try it and get a little creative in the process. This won’t be the best view to show what is going on for a given dealership (unless I put them into different graphs or emphasize only one or a couple at a time), but I can still give a sense of the highest, the lowest, and which fall generally in the middle when it comes to the most recent data by labeling in groups on the right-hand side of the graph.
STEP 3: Figure 5.2b shows my visual with these changes incorporated.
With Figure 5.2b, my audience can easily focus on the Regional Avg and also get a sense of the range and distribution over time across dealerships. If it’s important to have a more specific understanding of what’s happening for a given retailer, however, that’s more difficult. If I need to solve for that as well, rather than try to do more with this graph, I might augment it with another view of the data. We’ll look at that momentarily.
STEP 4: If I’m given a single slide to use as my communication vehicle, I’d want to put more words around everything to make sure it makes sense and attempt to answer the question of “So what?” I’d use titling and text together with my visuals—being conscious of white space and alignment—to create clear structure on the page. I’d also emphasize sparingly, both to create visual hierarchy and make the information scannable. This would help tie related elements together, easing the processing for my audience. See Figure 5.2c.
In Figure 5.2c, I added a second graph—horizontal bars showing how car sales across the various dealerships compare for the most recent time period. I’m making the assumption that this is the most relevant and that we don’t necessarily need the full view over time for each (we can see the highs and lows with relative ease on the left, but if it’s important to be able to distinguish the middle ones, that becomes impossible given the current design).
I’ve added more text around the graphs, both clear concise titling and descriptive text to help make what I’d like to highlight to my audience clear. I’ve used white space and alignment to create a two-sided layout. If we step back and consider how our audience is likely to process this information, they will probably start at the top left, read the slide title, then move downward and read “Overall decline in regional average” and see the black line below in the graph that depicts this. Then they’d typically move to the right-hand side, perhaps pausing on the “Marked variance by dealership” title or the blue and orange text. Finally, they’d look down to the right graph, see the black average tied to the graph on the left, as well as the blue and orange bars that are connected through similarity of color to the above words.
I did try out a second iteration of the left graph in Figure 5.2c that maintained consistent coloring of blue and orange for the top and bottom three dealerships as of Q3 2019 (those called out on the right). While I liked the consistency, I felt these competed too much for attention with the Regional Avg on the left, so I decided to use these colors sparingly on the right graph only.
The primary point here is to be thoughtful in the overall structure and design of your visuals and the pages that contain them. Don’t simply rely on tool defaults; once you make a graph, there is still more work to be done. When we design thoughtfully, we can create a better experience for our audience, improving the odds of successful communication.
The following example employs a two-sided structure similar to where we ended in Exercise 5.2. However, clear structure is not the only thing we need for success. Attention to detail is a hugely important aspect of creating effective visual design. Let’s look at another example and how attention to detail and thoughtful design choices can improve our visual communications.
Let’s assume you work for an on-demand print company that targets small businesses. One of the metrics you track is customer touchpoints—how many times someone at your organization interacts directly with a customer—both in aggregate and on a per-customer basis. There are three primary modes of connection: phone, chat, and email.
Your colleague has put together the following slide summarizing touchpoints over time and asked for your feedback. Spend a moment examining Figure 5.3a, then tackle the following.
STEP 1: What feedback would you give your colleague about the design of their slide related to attention to detail? Write down your thoughts. Focus on not only what you would recommend changing, but also why. Ground your feedback using design principles we have discussed.
STEP 2: Take a step back and think about how the data is designed: stacked bars on the left, table on the right, and additional numbers in the text. Are there changes you would make to the way this data is shown? How might you design the data in a way that is more intuitive for our audience? Write down your ideas.
STEP 3: Download the data and original visuals. Remake the slide, incorporating your feedback and ideas in the tool of your choice.
STEP 1: First, let me say that attention to detail is hugely important in our visual designs. Typically, the graph or the slide is the only part of the analytical process that our audience actually sees. Whether they should or not, people tend to assume things about the overall level of detail that was paid based on this piece that they can directly observe. So make your visuals and the pages that contain them imply good things about your overall work!
Related to attention to detail, I would concentrate my feedback on three areas: consistency, alignment, and intuitive axis labels. Let’s review each of these.
Consistency is an important aspect when it comes to attention to detail: be consistent in your approach unless it makes sense for some reason not to be. Changing design elements up randomly or otherwise introducing unnecessary inconsistency can be attention grabbing, distracting and looks sloppy. Specific things that catch my eye in this case are: inconsistent decimal points on y-axis labels of graph and in the bottom Email cell of the table. Also the way the dates are shown is inconsistent between the graph and the table, and not even consistent within the table!
When it comes to alignment, as we’ve discussed, centered text often looks messy. When it flows onto multiple lines, it creates jagged edges, as we see in the center-aligned statements above the graph and table. While I might preserve the centering of numbers in the table (if I were to keep the table; more on that shortly), I would be consistent in the vertical alignment. I would center consistently in that direction as well (currently the dates in the table are top aligned, while the numbers are center aligned vertically). Also, the overall elements on the page could be aligned a little better—the table isn’t directly under the line above it and the orange box on the far right could be sized to better fit the cells it’s meant to highlight.
My final main point of feedback on the current design would be in regards to intuitive axis labels in the graph. Currently, every fifth month is labeled on the x-axis. We can see why this was done: there isn’t sufficient space to label every point, particularly given the long format of the dates. One method is to label only some, though we should be thoughtful with what frequency we choose to label. Choose a frequency that will be intuitive based on the data being shown. For example, every seventh point labeled would make sense for daily data (since there are seven days in a week) or it could make sense to label by weeks instead of days. For monthly data, every third or sixth month would be more intuitive. If you have limited space with time on the x-axis, you could label by quarters or years. We could pull the years out as a supercategory and either abbreviate the months and arrange the text vertically, or just use the first letter of each month to maintain horizontal text. I’ll employ this latter method in my solution. There isn’t a single or preferred approach: choose axis labels that will be intuitive, helpful, and legible for your audience.
As additional points of feedback, I’d reduce redundancy by removing “Touchpoints” from each category label in the graph and also label the data directly so my audience doesn’t have to go back and forth between the legend and the graph to decipher the data. Color is also clearly something we can play with here, but I’ll reserve that for when I consider the overall design momentarily.
Figure 5.3b illustrates what my remake of the graph would look like incorporating the changes I’ve outlined.
STEP 2: When it comes to stepping back and designing the data in a way that makes sense, there are more sweeping changes I would recommend. Let’s shift next to how we might design the data to make sense.
Going back to Figure 5.3a, there are a lot of numbers between those called out in the titles, those added to the graph, and those in the table. We don’t need all of these. Let’s talk first about total number of touchpoints. This is referred to in the title and through text and numbers that have been added to the graph. If this information is critical, I could break it out on a separate slide and graph it (and would probably include more data than simply the two yearly numbers that are mentioned currently). Otherwise, I’d be apt to include the additional context as a sentence rather than clutter my graph with it.
Turning our attention to the table: this doesn’t add any new information. The data shown there is already graphed in the January points in the graph on the left. So rather than break it out separately, if these specific numbers are of interest, I’d recommend putting them on the graph directly with the data. In this case, I don’t think these numbers are critical. If we step back and think about the story, that will lead us to look at different views of the data, both to get a better understanding of where we want to focus and the story we can tell, as well as to figure out how to make that clear and easy for our audience.
Let’s focus on other ways we could visualize the data. One challenge with stacked bars is that we can really only compare the first data series at the bottom of the stack and the total (overall height of bars) with ease. If anything interesting is happening in a data series up the stack, it becomes quite difficult to see because those pieces are stacked on top of other pieces that are also changing. To allow for both of these comparisons with greater ease, I could unstack the bars and turn them into lines: see Figure 5.3c.
In Figure 5.3c, I unstacked the categories and graphed each type of touchpoint—Email, Phone, Chat—as lines. I added an additional line representing the Total. I also stripped color out of the graph entirely, so we can look at all of the data critically and determine where it might make sense to focus. We’ll add some color back in a later step.
When I look at this data, what jumps out at me—even more than with the stacked bars—is the apparent seasonality. When we want to clearly see seasonality (or in some cases, a lack of seasonality), it can work well to use a single year of months—for example, from January to December—for our x-axis, with a different line for each year. This change will result in a lot of lines if we do it for every category. With different data, we may need to split it into multiple graphs. However, here, given the spread of the data, we can make it work in a single graph. See Figure 5.4d.
In Figure 5.3d, I’ve changed the x-axis to January through December, plotting each year as its own line. Within each color grouping, the thin line represents 2018, the thick line represents 2019, and the circle points at the left represent our single month of data—January—for 2020. Notice that we see pretty consistent seasonality in Total touchpoints, with higher touchpoints per customer in January and December and relatively lower through the rest of the year. Don’t worry if you aren’t loving this graph—it’s an interim step to help get us to where we’re going next.
I’m going to assume that we’re standing in February 2020, since the most recent data point is January 2020. Given this, plus the shape of the data over the course of the year (higher at beginning and end, as mentioned, and lower in the middle), I am going to adjust my x-axis. Rather than the typical calendar year (January to December), I will change it to go from July to June to make it easier to see how recent months have compared year-over-year. In doing this, I’ll also eliminate some data, solve for the awkward single data points in 2020, and simplify my lines to “This Year” and “Last Year.” See Figure 5.3e.
With this view, I can make a couple of observations that didn’t jump out at me before. First, let’s pause on the Total: we see this year’s trend has followed last year’s closely. However, January touchpoints per customer are lower than last year. Moving downward, we see both Email and Phone touchpoints are trending lower this year compared to last year. Chat touchpoints, on the other hand, illustrate something different: Chat touchpoints have been consistently higher this year compared to last, with that difference increasing in January.
You may notice the varying decimal places on the labels in Figure 5.3e. I chose to round to one point past the decimal for Total and Email given the magnitude of the numbers. I took it out to an additional place past the decimal for Phone and Chat, both so that we can evaluate the small but potentially meaningful difference and so two points of varying heights wouldn’t be labeled with the same number (in this case 0.3), which could cause confusion.
STEP 3: Pulling this all together and putting words back around it, my final slide might look something like Figure 5.3f.
If I were talking through this information in a live setting, my slides would focus on the graph and I would build it piece by piece (we’ll look at examples of this in Chapters 6 and 7). However, if I have a single slide to get the information across—perhaps this is a slide that’s being incorporated into a broader deck that will be sent around—then I want to put all of the words around it so it makes sense. The words I’ve added are mostly descriptive; ideally we’d use this annotation to lend additional context, provide framing of whether what we are seeing is good, expected, and so on. I tied words to the data they describe through similarity of color. The result: when my audience reads the words, they know where to look for evidence in the data and vice versa. I used sparing color, relative size, and position on the page to create visual hierarchy and help make the information scannable.
By being thoughtful in all aspects of our design, we can make our data more easily consumable for our audience, helping ensure that our message comes across clearly.
Something we haven’t touched upon yet that can influence our design style when communicating with data is brand. Companies often go through great amounts of time and expense to create their branding: logos, colors, fonts, templates, and related style guidelines. Beyond being required to use this, there can be value in rolling branding into how you visualize data: it helps create a cohesive look and feel and can even add some personality into your data communications. Let’s practice applying branding to a graph!
We originally looked at the following graph in Exercise 3.1. Figure 5.4a shows market size over time for a given product. The storytelling with data typical look and feel has been applied. The font is Arial. Titles have been justified at upper left. Axis titles are in all caps. Most elements are in grey except sparing use of color to direct attention (orange for a negative callout and associated data point, brand blue for positive data point and corresponding comment).
Download the data and graph then complete the following.
STEP 1: Imagine you work for a brand similar to United Airlines and need to pull together an annual report that involves looking at market size. Start by doing some research: visit United’s website, search Google images, and browse related pics. Write down 10 adjectives that describe the brand. Recreate Figure 5.4a, rebranding with a style similar to United Airlines. Reflect on how this affects your choice of colors and font. How else might this brand influence changes in the design of this graph?
STEP 2: Let’s do this a second time. In this instance, you are an analyst at Coca Cola. Repeat the exercise, first by doing some research and making a list of words or feelings you’d associate with the brand. Then recreate this graph again, rebranding based on your research. What changes did you make to achieve this? How does red as a brand color play into your design?
STEP 1: Words that come to mind when I look at the United Airlines website and search Google for related images include: clean, classic, bold, blue, navigable, open, minimal, simple, serious, and structured. The logo has an intense dark blue background, with center-aligned, bold, white, capital letter text and sparing use of a lighter, more muted shade of blue. I can incorporate these feelings and elements into my design of the graph. See Figure 5.4b.
My main initial changes were to color and font. I used the dark and light blues throughout, with the exception of the graph axes: choosing black for axis titles and labels and grey for axis lines. The font I chose (Gill Sans) takes up a bit more space than Arial. This looked overly crowded with the text boxes above the data line. To remedy this, I moved the text boxes below the data and also reduced the y-axis maximum to shift the line upward, creating room below it to reposition the text boxes. I positioned the footnote below the graph.
I center-aligned most of the text (I played with left and right alignment of the large text boxes, and while I liked the structure of the clean edge that created, something about it didn’t feel fitting with the rest of the graph). The United logo and brand connote a feeling of clear organization to me, so I manifested that here by adding blue rectangles behind the title and footnote and also a blue border around the graph. I thickened the data line because I like how this balances out the bold title text. Even though the primary brand color is blue (similar to SWD), this rebranded graph feels quite different than the original Figure 5.4a as a result of these changes.
STEP 2: Next, let’s be inspired by the Coca Cola brand. I reviewed can and bottle labels, logos, and advertisements. Words I would associate with this brand include: red, silver, round, classic, bold, sweet, playful, international, diverse, and wet (there’s often condensation shown on the cans!). I observe a heavy use of red backgrounds, contrasting white text and sparing use of black. Text is typically center-aligned and frequently features a combination of bold all caps surrounded by slightly smaller non-bold all cap text. Words are used minimally. I’ll fold these components into my redesign. See Figure 5.4c.
One aspect of the Coca Cola brand that I chose not to incorporate is the cursive-like text in the Coca Cola logo. While this is fine for a logo, my priority for text related to the graph is legibility.
Text should be large enough to read and in a font that is easy to read. I opted for a sans serif font similar to the supporting text I saw on can and bottle labels (Montserrat, a free font that I downloaded). To incorporate some of the round feel that you get from the logo, I opted for a rounded (rather than rectangular) background shape.
Speaking of the background, the red background in Figure 5.4c is quite bold. This might be fine if it is the only graph we are looking at, or if graphs will be projected one by one on slides. If there will be multiple graphs on a single page or if I anticipate that my audience will want to print it, I may opt for a lighter “Diet Coke” version. See Figure 5.4d.
In Figure 5.4d, I opted for a light grey background, similar to the silver I saw incorporated into some of Coca Cola’s designs. With this lighter background, black stands out more, so I opted for a few more black elements compared to the original remake. I can use white, which fades to the background on grey (whereas it stood out a lot against red) for elements such as axis lines. I limited my use of brand red to the graph title and data.
Red as a brand color works well with grey and sparing use of black, and looks quite slick as we see in Figure 5.4d. When it comes to colors, there is a tendency to use red and green to denote bad and good or negative and positive, respectively. While I recommend against this due to considerations for colorblindness, I especially discourage it for organizations having red as a brand color. You want positive things associated with your brand, so if your brand color is red, don’t associate red with negative or bad things. One alternative in this circumstance can be to use red for good and black for bad. In the preceding graph, I’ve used red for general data and black for call outs (without connotation of bad or good), which is another option.
Stepping back and summing up: there can be value from rolling branding into how you communicate with data. If you work with client organizations, consider how you can undertake research similar to what we’ve done here and integrate your learnings into your designs. When it comes to you own organization’s brand, many companies have style guides that you can use to better understand the brand and what options you may have. Regard these not as annoying constraints, but rather as a lodestar that can inspire creativity and cohesiveness across your data communications.
One piece of advice I often give is to simply observe the examples of data visualization you encounter in the world around you. Pause to reflect: for the good ones, what works well that you can emulate in your own work? For the not-so-good ones, identify what pitfalls the creator fell into that you can avoid. Let’s do an exercise when it comes to the effective side of things.
Rather than simply pause and figure out what works well, we can go a step further and take the time to emulate the effective examples we identify, recreating them and learning how to achieve the aspects of effective designs in our tools. The level of attention to detail this process forces can help us be more thoughtful in our own work and sharpen our visual design skills and style. Let’s practice all of this!
First, identify a visual (graph or slide) someone else created that you believe is effective. This could be an example from a colleague at work, the media, storytellingwithdata.com, or elsewhere. After you’ve chosen an example, tackle the following.
STEP 1: Consider the four aspects of design we’ve discussed: (1) affordances, (2) aesthetics, (3) accessibility, and (4) acceptance. Judging from the visual you’ve chosen and making assumptions as needed for the purpose of the exercise—how did the creator account for each of these areas through the choices they made in their design? Write a few sentences describing how each of these four aspects of design were achieved.
STEP 2: Stepping back, why is it that the example you’ve chosen is effective? Are there specific elements of thoughtful design that make it work that you haven’t already described? How might you generally apply these learnings to your own work?
STEP 3: Is there anything about the example you’ve chosen that you believe is not ideal or that you would have done differently? Write a couple of sentences outlining your thoughts.
STEP 4: Recreate the visual you’ve identified in the tool of your choice. First, work to emulate it as closely as you can when it comes to the specifics (typography, color, and overall style).
STEP 5: Make another version that incorporates any of the aspects you outlined in Step 3 that you would have approached differently. Look at your visuals from Step 4 and Step 5 side by side. Which do you prefer and why?
It’s frequently a lot of little things that work together to create a great or not-so-great experience for our audience in the data communications we design. This means that small changes can have big impact in improving our visual designs. Let’s look at an example and also practice how these modifications can add up to help us take work from acceptable to exceptional.
Let’s say you work at an advertising agency and have been asked to assess a recent six-week ad campaign for a client. The data you are focusing on is incremental reach, which you measure “per 1,000 impressions.” You have a colleague who did a similar analysis for a different client recently, so rather than start from scratch, you’ve updated her visuals with your data as a starting point. Next, you want to edit and refine.
Figure 5.6 shows the visual you’ve created. Spend a couple of minutes to familiarize yourself with the details, then complete the following.
STEP 1: Pause first to consider what is working well. What do you like about the current view of the data?
STEP 2: A number of steps have been taken in Figure 5.6 to direct attention and help explain. Which are working well? Where and how might you adjust?
STEP 3: What clutter would you eliminate? What elements would you push to the background?
STEP 4: What other design choices made here do you question given the lessons in this chapter? What additional changes would you make?
STEP 5: Download the data and current graphs. Refine the visual by making the changes you’ve outlined in the steps above using the tool of your choice.
Imagine you work for the same on-demand printing company that we assumed in Exercise 5.3 when we looked at customer touchpoints data. How your company interacts with customers is one possibly interesting topic, as we saw. Another might be the competitive landscape for your products. As part of this latter area of focus, your colleague has been asked to pull together some data on your main competitors’ market share over time.
He comes to you with his slide—Figure 5.7—and asks for feedback.
Study Figure 5.7, then complete the following.
STEP 1: List 5 design improvements you would recommend making to this slide. Articulate not only what, but also why. How specifically will your ideas improve the design?
STEP 2: Download the data and execute the changes you’ve outlined in the tool of your choice.
STEP 3: Consider how you would present this material in a live meeting compared to something that has to be sent around as a stand-alone document. How would your approach change between these two instances? Write a few sentences to explain.
As we explored in exercise 5.4, there are ways that we can incorporate company or personal brand into how we communicate with data. This can be facilitated through choice of font, color, and other elements. In some cases, it may mean incorporating a logo or using a customized slide or graph template. Let’s practice how you can incorporate branding in a graph.
Suppose you work for a pet food manufacturing company. Look at the following graph, Figure 5.8, which depicts relative cat food sales over time (expressed in terms of % of total) for a given brand line, Lifestyle. Complete the following.
STEP 1: Identify two recognizable brands. They don’t have to be at all relevant to this example—these could be company brands or sports teams, for instance. It will be more fun and a better exercise if you pick two that are quite different from each other in terms of style. Research images related to the brand and list 10 adjectives that describe the look and feel of each. Remake this visual two times, incorporating branding components of each of these, respectively.
STEP 2: Take a step back and compare the two visuals you’ve created. How does each feel? Were you successful bringing to life the adjectives you outlined in Step 1? How can branding affect how we communicate with data generally? What are some pros and cons of this? Write a few sentences with your thoughts.
STEP 3: Consider your company or school’s brand. What descriptors would you associate with it? Remake the graph again, styling it accordingly. To take it a step further, integrate your branded graph into a slide, applying consistent branding to any elements you add (title, text, logos, and colors).
STEP 4: How would you generalize the components of brand we should think about when we visualize and communicate with data? What are the benefits of doing so? Are there scenarios where we may not want to be consistent with brand in our data communications? Write a few sentences outlining your thoughts.
When you look at a graph you made, it’s likely you know what you’re looking at: what to pay attention to, how to interpret it, and what to take away. But as we’ve discussed, this isn’t necessarily clear to our audience in the same way. Words used well can be a strategic tool for making our data comprehensible for our audience, answering questions before they arise, and helping them to draw the same conclusion that you have.
There are some words that have to be present: every graph needs a title and every axis needs a title. Exceptions to this will be rare (for example, if your x-axis reflects months, you probably don’t need to title it “months of the year”—you do, however, need to make it clear what year it is!). Make it your default to title axes directly so your audience doesn’t have to guess or make assumptions about that at which they are looking. Also don’t assume that people looking at the same data are going to walk away with the same conclusion. If there is a conclusion you want your audience to draw—which there should be when using data for explanatory purposes—state that in words. Use what we know about preattentive attributes to make those words stand out: make them big, make them bold, and put them in high priority places such as the top of the page.
Speaking of which—the top of the page (in Figure 5.9, “Words make data accessible!”) is precious real estate. It’s the first thing your audience encounters when they see your page or screen. Too often, we use this precious real estate for descriptive titles. Instead, use this for an active title; put your key takeaway there so your audience doesn’t miss it. This also works to set up what will follow on the rest of the page. (We’ll further explore and practice takeaway titling in Chapter 6.)
Also consider what is helpful to have present but doesn’t necessarily need to draw attention. For example, when showing data, it is often useful to have a footnote that lists details such as the data source, the time period represented (or time at which the data was extracted), assumptions, or methodology details. These are things that can help your audience interpret the data and lend credibility, as well as give you a reference in the event you need to replicate and create something similar in the future. It’s important, but doesn’t need to compete with other things for attention. This text can be smaller, grey, and in lower-priority places on the page, like the bottom.
After you’ve created your graph or slide, run through the following questions to help ensure you are using words wisely:
Affordances are aspects of our visual design that help our audience understand how to interact with the data we are communicating. We can draw attention to some components and push others to the background to create visual hierarchy and make our communications scannable. Want a quick test to see if you’ve done this well? Squint your eyes to see the overall impression of the chart. This changes your perception enough to get fresh eyes on a design. The most important elements should be the first things you see and the most prominent.
For more specific tips on how to achieve visual hierarchy, read through the following from SWD (paraphrased from Lidwell, Holden, and Butler’s Universal Principles of Design) for highlighting the important stuff and eliminating distractions. Determine how you can apply these to your next project!
Highlight the important stuff
Eliminate distractions
Many elements add up to create the overall experience our audience feels when faced with the visuals we create. Have you ever noticed how some designs feel easy and elegant, while others feel clunky and complicated? Paying close attention to details can help ensure the visuals we create are met with happiness by our audience. Here are some specific aspects of your visual design to consider to achieve this—the next time you create a graph or slide, read through and apply the following.
The following is adapted from Amy Cesal’s guest post on the SWD blog; you can read her full article, which includes a number of examples and links to additional resources, at storytellingwithdata.com under the title “accessible data viz is better data viz.”
Often, when we are creating charts and graphs, we think of ourselves as the ideal user. This is not only a problem because we know more about the data than the target user but also because other users might have a different set of constraints than we do.
Inclusive design principles and accessibility are important to take into consideration when designing data visualization because they help a broader audience understand your graphic. Designing with accessibility in mind can even help make your visualizations easier to understand for people without disabilities.
Being clear with text, distinctive labeling, and adding multiple ways to identify the point to your visuals will make it easier for people with impairments and those without to interpret your graphs. There are easy ways to add the principles of accessibility into your visual communications. Here are five simple ones:
These are just a few things you can do to help everyone easily comprehend the graphs that you create. You should strive to make sure that everyone—not just you or your ideal user—understands the point of the visualization. When you consider accessibility, you create a better product for all.
The next time you need to communicate with data, refer to and apply these tips!
People dislike change. This is a simple fact of human nature. In the scenario where we’ve always shown data in a certain way and people are attached to it—how do we convince them to do things differently? What should we do in general when met with resistance from our audience?
This is a change management process. In the same way that we considered our audience in the exercises in Chapter 1 and tried to understand what motivates them, we can do that here as well: in this situation our audience becomes those whose behavior we want to influence. First and foremost, when we want to convince our audience to be open to our designs, we need to do it in a way that works for them.
The wrong way to go about changing their minds sounds something like this, “I just read this book, and I learned that we’ve been doing it wrong; we should really be looking at it like this.” That might be easy, but it’s not so compelling or inspiring. So unless you’re the boss and people have to do what you say (even if that is the case, you should probably be more subtle in your approach!), you have to work to influence your stakeholders or colleagues to change.
Here are a few strategies from SWD—plus a couple of new ideas—that you can leverage for gaining acceptance in the design of your data visualization.
Reflect on whether any of the above can be employed in your situation to help you drive the change that you seek and the acceptance of your visual designs. In general, think about how you can set yourself up for success. Getting to know your audience—those you want to influence to accept your design—and what drives their behaviors can help. Think about not why you think they should change, but why they should want to. Make your approach work first and foremost for them. Refer back to Chapter 1 for exercises that will help you get to know your audience.
Also consider whether it’s a fight worth fighting. Don’t start with big battles. Start with low-hanging fruit and achieve small victories. Over time, you’ll build credibility so if and when you do want to make more sweeping changes, you’ll have earned your colleagues’ and audience’s respect and hopefully have an easier time making it happen!
Consider the following questions related to Chapter 5 lessons and exercises. Discuss with a partner or group.
18.119.131.235