In the midst of preparing for an internal training session on data visualization, I was thinking about why we don’t remember to do all the simple things to improve our analyses and reports so they can be more easily interpreted by others. The GEICO commercials came to mind as a perfect metaphor. If you can save 15% on insurance in 15 minutes, then why not improve your analysis by 15% in 15 minutes? I challenge all of us who share data with others to take 15 minutes to make a few improvements before we hit the send button for an analysis or report.
Good visualization and summaries mean it takes a lot less time for our clients to interpret our work, and gives them more time to focus on insights and how they can be used to improve their decisions. If we don’t make it easier for our audience, all of our efforts may be wasted when people misinterpret our results or simply give up.
What can you do in 15 minutes? Here’s my list of 7 ways to improve data summaries and visualization (admittedly you may not be able to do ALL of these in 15 minutes…).
- When you send off a report, tell them what it is and why they should look at it. “Here’s the analysis” is not likely to spark my interest to look inside. Entice your audience in your email. If it’s a longer analysis, add an executive summary. In an Excel file, add a description tab up front that describes all the subsequent tabs, fields, definitions, etc.
- Label, label, label, and label in English. Too often I see abbreviations and terms which may make sense to the author and three other people, but not to a broader group. Make sure your output is clear enough to be shared upwards or across the organization.
- Make charts as simple as possible. For example, abbreviate months; eliminate decimal places and round up to thousands or millions where appropriate; remove extraneous lines; don’t go crazy with color, 3-D or other fancy formatting.
- It you think they might print it, then make it printable. One of my biggest pet peeves? I open an Excel file and print it only to find that it prints on 4 pages because it is 1 column wider than a page and 1 row longer than a page.
- While we’re on the topic of printing, use headers and footers. A header with the filename inserted means people can easily reference the source file at a later date. If it’s more than one page, add a footer that says “Page <page number> of <number of pages>.”
- Make large tables easier to digest by locking header rows and/or columns, adding filters, and sorting the table by a relevant column. Another option is to add a summary table that provides counts and sums by key categories.
- Finally, consider incorporating a Buddy Check process. Per Wikipedia, this is originally attributed to SCUBA divers, who check each others' equipment prior to a dive. Why not have a peer take a quick look at your analysis before you distribute it? When we’re so close to the analysis, we sometimes overlook the obvious.
I’d be interested to know whether your list of improvements is different from mine. Add a comment on your favorite tips to improve your output.
About the Author: