When my graduate advisor gave me feedback on the first chapter of my master’s thesis, which was about 25 pages long, he circled a paragraph in the middle of page 22 with a big red marker and wrote ‘START HERE!’ in big letters over the rest of the page.
Ugh . . . so I copied that frail, little paragraph into a new, giant, blank document and started over, from scratch.
I would eventually learn that my error was not that what I wrote before page 22 was wrong, but that it was very difficult to make sense of. The chapter followed the path of my own learning and thinking process, and not what the reader needed.
That is the key to presenting data well. First, you learn what is to be learned so you understand the material (my first draft). Second, you need to PRIORITIZE and SHAPE that learning for other people to digest (my second draft, which my advisor liked much more). You need to shift from learner to teacher.
In many of our presentations, we do not make this shift and so we tend to wander through all the fits and starts of our project or our analysis, and hit the punchline somewhere on the second half of slide 22. No one remembers the punchline, and many people get confused.
There is a better way. We tend to feel that more data is always better, but it rarely is when it comes to presentations. Technical aspects of visualizing data have been covered earlier in this blog (in posts on presenting data generally, presenting data in tables, and presenting data in graphs). This post will look at the process of presenting data in the context of a larger presentation.
Start by getting clear about the key message of your presentation. Write it out in sentence form and keep it in front of you while you work. Outline the presentation on paper with a pencil and think hard how that one central message will be grasped. Data presentations are usually only part of the story.
Those of us who have done research have learned an organizational for presentations that has a parallel in other types of projects. It is a good place to start.
(1) Lay out the question or the problem statement.
(2) Spend one or two slides on what the work consisted of, special methods you had to employ, information about the data and its problems, or collaborations you relied on.
(3) Present your findings/data/proof/digging. Start with the most important graphic (the one that tells the story) and spend a couple more data slides adding detail and nuance (but only if it is importance to your question or problem).
(4) Summarize your conclusions, make your recommendation, ask for feedback or direction, or summarize the next steps.
That’s it. You should be able to do most projects in seven slides or even less. And the presentation should deliver your core message without confusion.
Here are some basic guidelines for presenting your data that will get the audience on your side.
(1) Give the audience TIME to digest each visualization. As you present, describe what is on the y- and x-axes, what the bars or lines represent, if there are special ways you transformed the data, excluded cases, etc.
Your first visualization gives you a chance to gauge your audience and decide on the pace for the rest of the talk. Look to see if heads are nodding or eyes are squinting in confusion. Sometimes I decide at that point to skip things that might be too in the weeds for the audience once I do this quick assessment.
(2) Include only one graphic per page, unless there is something to be learned in comparison. There are GREAT examples of presentations with multiple graphics on a page. But be aware that doing this well takes time, skill, and knowing your audience. It will take your audience longer to understand the material, too. The different graphics should work together to tell one story. Here is an excellent example of a page with multiple visualizations (compliments of Ann Emery’s wonderful blog) –
(3) Avoid animations unless they walk the audience through the graphic in a way that aids their understanding. I get frustrated with animations that slow down my ability to grasp the slide. Presentations are about communicating something you know to people who do not know it. Focus on doing that really well.
(4) Label everything, but be parsimonious. Try to reduce the words, lines and numbers displayed to a minimum. It will take the audience a few minutes to make sense of your graphic, longer if there are more words and numbers. Make it easy for them to get to your message quickly. I like the work of Ann Emery, mentioned above, and on her blog, she shows how she transforms clunky, out-of-the box graphics into easy-to-digest visualizations that actually include more information. There is a lot to learn from her! Here is an example from her blog:
(5) Document your data in footnotes. In healthcare environments, audience members are appropriately sophisticated about how data definitions can shift the conclusions of an analysis. Be prepared for questions on this, which your footnotes should answer. Here is a very simple way to do that, again, from Ann Emery:
(6) If you are displaying several, similar graphs, be exacting and consistent with formatting so that the eye just needs to focus on the data points as you flip from slide to slide. The audience will grasp the conclusions faster. I do this by creating one slide just exactly as I want it and then copying it into the next slide and just changing the data. That way, the audience does not need to relearn the y- and x-axes. They can just focus on the data (remember to keep your axes consistent whenever possible).
(7) Use appendices for supporting material that is not essential. If it turns out you need this information, you can quickly flip to these slides. But short and to the point will almost always be enough.
It is never a bad idea to show your slides to a colleague before presenting. I do that all of the time. And usually my colleagues catch an error or a point of confusion that I did not see. Asking for someone to review my work is one of the best things I was ever taught.
Comments will be moderated.