6 Comments

  • Kim says:

    Nice 🙂

  • Ione Farrar says:

    This is beautiful, Ann. Which version of the 3 would you use in a report? I like the way you labeled the years of interest on the horizontal axis. I’ve been trying to reduce labels by selecting the interval in format axis, but it tends to omit the label for the final year. Did you include all the years on the horizontal axis and delete selectively?

    • Ann K. Emery says:

      Hi Ione,
      I’d probably use the third one (with both of the milestone years labeled) in a report.
      To reduce labels–if you want to show, say, every other year (or every third year, or every fourth year, at regular intervals), then you would adjust the Interval in the Format Axis screen.
      If you want to show labels at irregular intervals, like I did here, then you would just delete a few labels in your spreadsheet. I added a free download above so that you can explore the spreadsheet and see how I did it.
      Good questions!
      Ann

    • Ann K. Emery says:

      Oh, one more thing! In a report, I’d also change the title a bit.. Something like “AIDS Diagnoses and Deaths are Declining” and a subtitle that noted the peak, like “AIDS diagnoses peaked in 1993 and have been declining over the past two decades. Deaths from AIDS peaked in 1996 and have also been declining over the past two decades.”

  • Leave a Reply to suegriffey Cancel reply

    Your email address will not be published. Required fields are marked *

    You want to display a lot of historical data–great! But sometimes we have so many points in time that our graph’s labels get smooshy. In this post, I’ll show you a before/after data visualization makeover in which we selectively labeled a few key milestones in order to tell our story (and make the graph more legible).

    Before

    Here’s an epidemiologist’s before slide.

    There are a lot of things going well in this slide already. For example, the graph already contained direct labels, the words classifications and deaths, rather than having a separate legend. Direct labels save viewers time because viewers don’t have to zig-zag their eyes across the slide and hunt for information. Direct labels are also better for people with colorblindness and for grayscale printing scenarios.

    Another feature that was already working was the branding. This epidemiologist’s center used these exact shades of turquoise, blue, green, purple, red, and yellow across all publications. They also used Calibri in all publications. We didn’t need to change the colors or fonts.

    Our priority on this graph was to make the line graph’s years easier to read. We wanted to avoid diagonal text, which takes longer to read than horizontal text. We also wanted to avoid the 9- point font. The years were barely legible for people sitting in the front row, let alone the back row. If nobody can read your graph, why bother?

    Here’s an epidemiologist’s before slide. It has a lot of positives but we need to make it easier to read.

    After

    Here’s our makeover. We:

    • Transformed the title into a subtitle;
    • Wrote a new title, which is purposefully short for a live conference presentation (we didn’t want the audience to be reading a lengthy title while the epidemiologist was speaking);
    • Deleted the vertical label on the y-axis;
    • Deleted the horizontal label on the x-axis;
    • Abbreviated the years (1985 to ’85) to save a little extra space;
    • Only labeled a few milestone years;
    • Added a few numeric labels along the line (e.g., 77,173 diagnoses in 1993);
    • Increased all font sizes to 18 (and a 32 title); and
    • Removed the logo.

    Two techniques fixed our smushed labels. First, abbreviating years—from 1985 to ’85—saves some space. Second, only labeling a few years—’85, ’93, ’96, and ’14—saves space and focuses attention on the milestones of your choosing.

    We intentionally drew the audience’s attention to ’93 and ’96 to tell our story and match the epidemiologist’s talking points, as described below.

    Here’s an epidemiologist’s after slide with changes made to make it easier to read.

    Storyboarding for Live Presentations

    Here’s how I envision the epidemiologist presenting the information. She would storyboard the information across three separate slides.

    First, she would present this slide, and say: “We track how many people in the U.S. are diagnosed with and die from AIDS. This graph shows how many people were diagnosed and died between 1985 and 2014. In 1985, nearly 12,000 people a year were diagnosed with AIDS. In 2014, closer to 20,000 people a year are diagnosed with AIDS. And of course, there have been some ups and downs over the years, which we’ll talk about next.”

    This graph shows how many people were diagnosed and died between 1985 and 2014. In 1985, nearly 12,000 people a year were diagnosed with AIDS.

    Second, she would present this slide, and say: “I’m sure you noticed this peak right away. Seventy-seven thousand people were diagnosed in 1993. In 1993, there was a peak in diagnoses, probably because we expanded our definition.”

    This graph has a peak that shows seventy-seven thousand people were diagnosed in 1993.

    Third, she would present this slide, and say: “1996 was another milestone year. Beginning in 1996, we’ve seen fewer deaths, probably because of the success of highly active antiretroviral therapies.”

    This graph shows that 1996 was another milestone year. Beginning in 1996, we’ve seen fewer deaths, probably because of the success of highly active antiretroviral therapies

    In our makeover, the horizontal axis is actually legible, hooray!

    We also transformed a single slide into three slides. The epidemiologist’s speaking points would be identical in both cases. Her pace would also be identical in both cases. She doesn’t need to talk any faster just because she’s got three slides instead of one. Only the visuals would change. Adding more slides doesn’t make my presentation any longer. Adding more slides guarantees that her audience’s attention is focused on the right part of the graph at the right time.

    Have you encountered smooshed line graph labels in your own projects? How have you fixed this common problem?

    Bonus: Download the Spreadsheet

    Want to see how I selectively labeled a few key points on the horizontal x-axis? Download the spreadsheet.


    Download the Spreadsheet ($0)

    More about Ann K. Emery
    Ann K. Emery is a sought-after speaker who is determined to get your data out of spreadsheets and into stakeholders’ hands. Each year, she leads more than 100 workshops, webinars, and keynotes for thousands of people around the globe. Her design consultancy also overhauls graphs, publications, and slideshows with the goal of making technical information easier to understand for non-technical audiences.

    6 Comments

  • Kim says:

    Nice 🙂

  • Ione Farrar says:

    This is beautiful, Ann. Which version of the 3 would you use in a report? I like the way you labeled the years of interest on the horizontal axis. I’ve been trying to reduce labels by selecting the interval in format axis, but it tends to omit the label for the final year. Did you include all the years on the horizontal axis and delete selectively?

    • Ann K. Emery says:

      Hi Ione,
      I’d probably use the third one (with both of the milestone years labeled) in a report.
      To reduce labels–if you want to show, say, every other year (or every third year, or every fourth year, at regular intervals), then you would adjust the Interval in the Format Axis screen.
      If you want to show labels at irregular intervals, like I did here, then you would just delete a few labels in your spreadsheet. I added a free download above so that you can explore the spreadsheet and see how I did it.
      Good questions!
      Ann

    • Ann K. Emery says:

      Oh, one more thing! In a report, I’d also change the title a bit.. Something like “AIDS Diagnoses and Deaths are Declining” and a subtitle that noted the peak, like “AIDS diagnoses peaked in 1993 and have been declining over the past two decades. Deaths from AIDS peaked in 1996 and have also been declining over the past two decades.”

  • Leave a Reply to suegriffey Cancel reply

    Your email address will not be published. Required fields are marked *

    You Might Like

    Our complimentary mini course for beginners to dataviz. Takes 45 minutes to complete.

    Enroll

    Inside our flagship dataviz course, you’ll learn software-agnostic skills that can (and should!) be applied to every software program. You’ll customize graphs for your audience, go beyond bar charts, and use accessible colors and text.

    Enroll

    Subscribe

    Not another fluffy newsletter. Get actionable tips, videos and strategies from Ann in your inbox.