Colors direct our eye movements, and therefore our brains and attention.

It’s up to you: will you help or hinder your reader’s understanding?

Step 1: Start with Your Brand Colors

Otherwise, your graphs, slides, and dashboards will be Frankensteined.

I’ve written about brand colors and brand presents in other posts.

Some of those resources include:

• Examples of organizations’ brand colors used in graphs
• How to read color codes in style guides
• How to enter your custom color codes in Excel

Step 2: Do Your Accessibility Testing

I’ve written about colorblindness, color contrast, grayscale printing in other posts.

Some of those resources include:

• An official color contrast test
• An official colorblindness and grayscale printing test

Then, your accessibility testing “results” should go inside your organization’s Dataviz Style Guide.

Step 3: Apply Those Brand Colors According to the Data & Variables

Now, it’s time to apply those branding colors to ensure that your graph is intuitive.

Look at your graph: Is your variable binary, sequential, diverging, or categorical?

Or, do you want to tell a story with a dark-light contrast?

Binary Variables Get Binary Color Schemes

Binary variables include yes/no data, such as:

• yes/no survey questions
• people who speak Portuguese as their primary language vs. people who don’t
• people who own a home vs. people who don’t
• people who graduated from program on time vs. people who didn’t
• people diagnosed with an illness vs. people who don’t have it

For binary variables, choose one brand color. The “presence” of the attribute gets the darker color, and the “absence” of the attribute gets the lighter color.

Here’s an example:

Sequential Variables Get Sequential Color Schemes

a.k.a. ordinal

Sequential variables have a natural order.

Examples include:

• age ranges (5-9 year olds, 10-14 year olds, and 15-19 year olds)
• income levels
• highest educational level completed (some high school, high school diploma, some college, etc.)
• years (Year 1, Year 2, and Year 3 of a project)
• semesters (fall, spring, fall, spring…)
• cohorts (first cohort of participants, second cohort, etc.)

For sequential variables, choose one brand color, and use a light-dark gradation of that color.

Here’s an example:

Categorical Variables Get Categorical Color Schemes

a.k.a. nominal

Categorical variables include:

• race/ethnicity (African American, Asian, Hispanic/Latin@, White, etc.)
• gender (male, female, nonbinary, genderfluid, etc.)
• chapters of a report
• sections of a presentation
• categories of a dashboard

For categorical variables, use a different brand color for each category.

Here’s an example:

Diverging Variables Get Diverging Color Schemes

Diverging variables are opposites.

Examples include:

• agree/disagree scales on surveys
• changes over time (e.g., “50 percent decrease” or “70 percent increase”)

For diverging variables, choose two brand colors, and place the darkest shades on the poles.

Here’s an example:

Combining these Techniques

In most real-life projects, we need to combine these color techniques.

In this map makeover, for example, we needed to:

• use brand colors, not software defaults;
• use two brand colors, one for each category; and
• apply a dark-light gradation to each map, because these are ordinal variables.

In this population pyramid makeover, we needed to:

• use two brand colors, one for each timeframe, and
• apply a dark-light storytelling emphasis to each pyramid.

Your Turn

What types of color questions do you have? Comment below..

The post How to Apply Your Brand Colors in Dataviz (Ordinal, Diverging, Categorical, and More) appeared first on Depict Data Studio.

Before: Just the Survey Questions; No Graphs Yet!

A couple weeks ago I spoke to Harvard University graduate students about visualizing survey results. This not at all to very survey scale is quite common in my research circles so I’m sharing our ideas with all of you, too.

Here are some of the survey questions that were asked before and after program participation. Don’t worry, the online survey was formatted much more beautifully than this screenshot from the Microsoft Word version of the survey!

After: Stacked Columns for the Ordinal Survey Scales

I’m going to show you two makeovers. In both makeovers, I used stacked column charts to display these ordinal survey scales.

Why did I choose vertical column charts instead of horizontal bar charts? I displayed the patterns over time from left to right across the page. Before results go on the left and after results go on the right, so they get vertical columns (not horizontal rows).

Why did I display percentages instead of numbers? There were more than 100 survey responses, so I converted the raw numbers into percentages. When there are fewer than 100 responses, I display the raw numbers. It’s super confusing to talk about 16.56% of 14 people.

Why did I use three colors? I suggest that you color-code by category because it’s a great strategy for breaking up oceans of information into manageable chunks.

Why did I use different shades of each color? The very knowledgeable to not at all knowledgeable scale is ordinal so I used darker and lighter versions of each hue to correspond to the amounts of knowledge.

Why did I write titles and subtitles? I’m a visual person and prefer reading graphs over paragraphs but some viewers will prefer reading paragraphs over graphs. One of the most common mistakes I see among novices is that they focus so much on creating the graph that they forget about the paragraphs. Your viewers will benefit from having both.

Makeover 1: The Traditional Data Visualization Approach

Before you create any data visualizations, I suggest doing some upfront planning with your colleagues. You’ll want to discuss a few thought-starter questions in advance. For example, who’s your audience? Are you designing your visuals for a technical or a non-technical data audience? Are you making a slidedeck, a handout, a technical report, or some other dissemination format altogether?

Here’s the most important planning consideration in data visualization projects: Are your viewers expecting a story?

Sometimes your viewers will expect a traditional data visualization approach, in which they’ll (hopefully) read between the lines and uncover a takeaway message for themselves.

Other times, your viewers will expect a storytelling data visualization approach, in which you use a dark/light color contrast and explicit text to uncover the takeaway message for them.

Here’s the first makeover, which uses a traditional data visualization approach:

Makeover 2: The Storytelling Approach to Data Visualization

Stacked bar charts are one of the most common ways to display survey results because surveys often include scales like this one.

But, we have to be careful because one page with two points in time, three survey questions, and five options per survey question can get cluttered, fast!

In this version of the handout, I used saturation to guide the viewer’s eyes towards the very and moderate categories. In other words, those categories are dark while the other categories are light. This dark/light color contrast is called a preattentive attribute, and it’s one of the easiest strategies for telling a story with data. Your viewers don’t have to think about it or waste any of their precious mental energy. Their eyeballs are instantly drawn to the darker parts of the page.

In your project, you may choose to draw attention to the not at all knowledgeable category. There are several correct ways to guide eyes with saturation. Your job is to anticipate what your viewers will find most useful. Just use your best professional judgment.

Here’s the second makeover, which uses a storytelling data visualization approach:

Learn More about Visualizing Survey Results

In this blog post, we looked at visualizing how participants’ knowledge about historical events increased after participating in an educational program at a museum.

Here are additional resources that show you how to visualize survey results in Microsoft Excel:

• Agree/disagree scales
• Check-all-that-apply questions
• Demographic data 
• Workshop satisfaction survey questions

Bonus: Download the Materials

Want to explore these graphs in more detail? Download the Microsoft Excel spreadsheet and the Microsoft Word document that I used to create these handouts.

Purchase the templates

The post How to Visualize Pre/Post Survey Results in Microsoft Excel appeared first on Depict Data Studio.

Use Horizontal Bar Charts When You’re Graphing Nominal Variables

Nominal variables—favorite ice cream flavors, types of organizations where conference attendees are employed—can be arranged in any order. I arrange nominal variables in a list from top to bottom, which means their bars are horizontal.  You can sort nominal variables from greatest to least or from least to greatest. Both are correct. My advice: Sort the data so that the item that warrants attention is displayed first—a high number deserving a celebration, or a lower-than-hoped-for number that needs to be turned around.

Use horizontal bar charts to display nominal variables like favorite ice cream flavors or employment settings.

Use Vertical Column Charts When You’re Graphing Ordinal Variables

Ordinal variables follow a natural progression—an order. Another name for ordinal variables is sequential variables because the subcategories have a natural sequence. I arrange ordinal categories from left to right so my viewers can view the sequence across the page, which means their bars are vertical.

Use vertical column charts to display ordinal variables like age ranges, salary ranges, and even cohorts or graduating classes (e.g., the percentage of students from each graduating class who achieved x outcome).

Real-Life Examples

Okay, you’ve seen some fictional, generic examples. Let’s look at a few examples of horizontal bar charts and vertical column charts from real projects.

Compare Age Ranges with Vertical Histograms

While working with a museum, we wanted to display some key demographic characteristics about people who had responded to the museum’s survey. Age ranges are ordinal, so we used vertical column charts to visualize how many people fell into each age bracket.

While working with public health researchers, we needed to visualize how many males and females in each age range were diagnosed with x disease. We built stacked column charts with vertical columns. We also used a population pyramid to compare the distributions of males and females.

Compare Cohorts of Survey Respondents with Vertical Column Charts

In the State of Grantseeking project, an organization called GrantStation sends out surveys a couple times a year, in the spring and fall. We wanted to compare how different iterations or cohorts of survey takers responded. In other words, we wanted to see whether people who responded to the spring 2016 survey were any different from people who responded to the fall 2016 survey or the spring 2017 survey. Cohorts are ordinal, so we used vertical stacked column charts to depict the proportion of survey respondents who are employed in nonprofits.

Compare Patterns Over Time with Vertical Charts

While working with a hospital’s analytics team, we needed to display the proportion of hospital procedures (“Product A” and “Product B”) that were performed by this hospital (“the ABC Org”). We wanted to compare two points in time, 2012 and 2016. Time is ordinal, so we used vertical columns.

Visualize Agree-Disagree Scales with Horizontal Stacked Bars

While working with a museum, we needed to display how many people agreed or disagreed with each statement on a survey. Agree/disagree scales are ordinal. (Well, technically, agree/disagree scales are a special type of ordinal variable called a diverging variable.) We wanted our readers to see the natural progression from agree to disagree, so we used horizontal bars to show that progression from left to right across the page.

Compare Before-After Survey Responses with Vertical Column Charts

In this example, we needed to make a one-page handout that summarized whether program participants were more knowledgeable about New England history, Martha’s Vineyard, or whaling after completing an educational program at a museum. The museum administered the survey twice, at the beginning of the program (pre) and at the end of the program (post). Pre-post comparisons are comparisons over time… and time is an ordinal variable… so we selected vertical columns. But wait! There’s more! We were graphing two sets of ordinal variables: 1) the timeframe (pre or post survey administration) and 2) the scaled survey responses (very, moderate, somewhat, slightly, or not at all). We chose to give the timeframe more weight than the scaled survey responses. In other words, we decided that the pre-post timeframe was the most important, and used vertical columns first and foremost. Then, within those columns, we displayed the ordinal survey responses.

These are guidelines, not rules. If you can explain the logic for going against this guidance, then your graph is probably going to be alright. My goal is to develop critical thinking masterminds, not robots.

Bonus: Download the Materials

Download the Excel spreadsheet that I used to make the generic bar and column charts at the top of this article.

Download the template.

The post When to Use Horizontal Bar Charts vs. Vertical Column Charts appeared first on Depict Data Studio.

Rules for Using Pie Charts and Donut Charts

Here are my guidelines for using pie and donut charts. Pie charts are okay when they:

1. are well-formatted. No 3D, exploding slices, leader lines, or legends.
2. display nominal variables. Ordinal variables don’t belong in a pie chart.
3. add to 100%. I’ve seen pies that only add to 90% because the designer deleted the “other” category and forgot to recalculate the new percentages.
4. contain positive numbers. I’ve seen designers place a mix of positive and negative numbers inside the same pie chart, which doesn’t make any sense.
5. display a single point in time. Patterns over time belong in a time series graph, like a slope chart, line chart, or dot plot.
6. only have two or three slices. Four slices is pushing it.
7. are displayed individually. Only show one pie chart at a time. No small multiples pies. Comparisons across multiple pies are time-consuming.

Finally, while I don’t consider this to be a strict guideline, pie charts tend to be easiest to read with common fractions, like a one-fourth vs. three-fourths pie or a one-third vs. two-thirds pie.

Circumstances When Pie Charts Are Okay

Given these guidelines, I use pie charts to show:

• male/female/etc. gender categories;
• yes/no survey responses;
• students who graduated high school on time vs. didn’t graduate high school on time;
• adults who live in single-family homes vs. adults who live in other housing types; and/or
• other binary data.

Alternatives to Pie Charts

Let’s tackle these pie charts! It’s not sufficient to tell you to avoid pie charts. You need to know what to do instead. Here are some pie chart makeovers that are inspired by my real projects.

If Your Pie Chart is Poorly Formatted… Then Format It!

The chart on the lower left is poorly formatted. This one is 3D, so the slices look larger or smaller than they really are… and, it’s exploding, which is distracting for viewers… and instead of the percentages being right on top of the pie slices, now there’s a tiny legend down below the pie, which means our viewers would have to zig-zag their eyes around the slide to tell which slice is which.

The final sin in this poorly-formatted pie chart is that there are leader lines, those gray lines connecting the 25% and 75% to their corresponding slices. Plus, So much ink is on the page, yet so little is actually focused on the data.

The well-formatted pie chart on the lower right is fair game. Gender is nominal or categorical, so that works. We’re only showing a single point in time, so that works too. And we’ve only got two different slices.

If You’ve Got Ordinal Data… Then Use a Stacked Bar Chart or a Column Chart

Ordinal or sequential variables have a natural order, like responses to a survey that go from strongly agree to agree to disagree to strongly disagree.

In this case, you’d swap out your pie chart and use a stacked bar chart instead, so that viewers can tell which category is at which end of the spectrum – the agrees on one side and the disagrees on the other side.

Another type of ordinal or sequential variable is age ranges. Histograms are a great alternative to pie charts when you’ve got ordinal or sequential groupings. Let your audience read across the screen from left to right (i.e., from lowest to highest).

If You’ve Got Negative Numbers…Then Use a Deviation Chart

I mentioned that pie charts are only for positive numbers, not negative numbers. Sometimes we have negative numbers when we’re dealing with changes over time.

In this example, we’re looking at four products and whether they increased or decreased in sales compared to the previous quarter. For example, Product A’s sales decreased 20% compared to the prior quarter while Product B’s sales improved 40% compared to the prior quarter.

Instead of a pie chart, we’d use a column chart or bar chart. In your software program, the negative numbers will automatically flip in the opposite direction of your positive numbers. The axis line runs across the middle at 0% and we can see which products went down (like Product A) and which products went up (like Products B, C, and D).

If You’ve Got Patterns Over Time… Then Use a Time Series Chart

What if you have time series data, that is, patterns over time? Maybe you’re trying to show data for each Quarter – Quarter 1, Quarter 2, Quarter 3, and Quarter 4 – or for each month, or for each year in the grant cycle.

Swap out your pie chart and use a line chart instead. You want viewers to see the beginning point – Quarter 1 – over to the end point – which is Quarter 4.

If You’ve Got More than Two or Three Categories… Then Use a Bar Chart

Pie charts are easiest to read with only two or three slices.

What if you have lots of different slices, like favorite ice cream flavors? This pie chart has too many slices – vanilla, chocolate, strawberry, mint, and cookie dough. It’s too hard for our brains to compare the slices to each other.

Swap out the pie chart for a bar chart and order the bars from greatest to least (or least to greatest). Chocolate would be listed first because it’s the most popular, and cookie dough would be listed last because it’s the least popular.

If You’re Tempted to Display More than One Pie at a Time… Then Use a Grouping of Stacked Bar Charts

What if you want to compare several companies, organizations, outcomes, etc. all at once? Pie charts are hard enough to read. Our brains don’t do well deciphering the angles, area, or circumference of circles. Two or three or four different pie charts can be understood, but with way too much mental energy.

In this example, we’re asking our viewers to look first at the 20% angle, and then at the 40% angle, and then their eyes have to zig-zag to the 60% angle, and then their eyes have to zig-zag over to the 80% angle. So. Much. Work.

In this case, you’d swap your small multiples pie chart for a small multiples stacked bar chart. The part-to-whole pattern is still there, but now our viewers’ eyes only have to make a single, diagonal swooping motion down the page to compare all four companies at once. Less energy required for reading, more energy reserved for making decisions based on that data.

Join the Conversation

Have you seen real-life pie charts or donut charts that are in desperate need of a makeover? Comment below and include a link to the example, and I may even include it in a future pie chart makeover blog post.

The post When Pie Charts Are Okay (Seriously): Guidelines for Using Pie and Donut Charts appeared first on Depict Data Studio.