Data Viz Checklist

A complete data visualization tells a whole story, not half a story. This checklist will help ensure that your story is complete. You should also review ProPublica’s Style Guide, which quite thorough and thoughtful.

The Data

[ ] Question the context Do you understand the context this data was collected in? Your captions should reflect that context.

[ ] Know and normalize your numbers. Did you say “average” when you really mean “median”? Are your numbers normalized so that you’re comparing apples and apples? If you need to combine some columns or drop others, are your choices rational (and not simply convenient?)

[ ] Talk to someone who understands the numbers Your story isn’t complete unless you’ve actually spoken with at least one source who has some familiarity with the numbers that you’re looking at.

[ ] Link to the raw data Newsrooms will take different positions on this one, but for class assignments, you must include a link to both the source of your data (the site you downloaded it from) and the data you used to generate the visualization. If those are one and the same, great. But if you had to clean the data or manipulate it to make your visualization, share that cleaned data with your readers.

[ ] Cite your sources Who did you get the data from, and when? Provenance matters: if the ACLU has compiled a years worth of quarterly NYPD reports and published them as a single free standing data set, did you get the data from NYPD or ACLU?

The Visualization

[ ] Lighten the cognitive load Don’t make your readers work. Every visual translation is a layer between content and reader. If readers have to flip between tabs to make comparisons, or refer to too many legends, they’re working too hard understand your story.

[ ] Choose visual encoding wisely What do properties like position, length, angle, area, hue, saturation, or brightness represent in your visualization?

[ ] Focus on the data Are you maximizing the data-to-pixel ratio? Use pixels wisely by avoiding chart chunk and decoration that doesn’t add to a reader’s understanding of the data.

The Text

[ ] Add captions that help the story along A good caption expands on your visualization without describing it. I could say “this checklist includes the ingredients for a complete data visualization” but you already know you’re looking at a checklist. So use the caption to add something.

[ ] Hyperbole is the death of a story. Make sure you can back up every statement of fact in your text. If you need to introduce a story with a line like “More Americans than ever are buying pants these days” make sure you know it’s true, even if your story is not about the growth in the pants market at all. Always be accurate.

[ ] Never “click here” — if you have to say “click here” your project is missing something. Show it, don’t say it.

[ ] Put your reporting in context with links. You aren’t reporting in a vacuum — highlight the reports and reporting that put your story in context.

[ ] Edit for redundancy Every word counts. If you find that you’re using the same phrase again and again, that’s a sign that you need to rework your text. Redundant:

Total Arrests in 2008: 15
Total Arrests in 2009: 39
Total Arrests in 2010: 422
Total Arrests in 2011: 3

Much better:

Total arrests:
2008: 15
2009: 39
2010: 422
2011: 3

[ ] Edit for brevity by looking for filler text and rewrite it out. Phrases like “here you can see” or “in this chart” or “this graph shows” just fill up precious space. Tell the reader why the numbers are interesting (and do it without saying “these numbers are interesting because …”)

[ ] Edit for grammar and style Take one last pass and check your grammar.