On the charts

December 21, 2021 6:30 AM

Practical advice on visualizing scientific data, from Nature Methods
posted by Gyan (5 comments total)

This series has long been a favorite of mine.
posted by grouse at 9:51 AM on December 21, 2021

Also see The Science of Visual Data Communication: What Works.
posted by NotLost at 11:15 AM on December 21, 2021 [3 favorites]

If someone is going to tell me about good graphic design, it would nice if their graphics were actually made with good graphic design. Teeny tiny pictures and text, just don’t cut it. It’s like they just shrunk some PowerPoint slides for their article. I genuflect in the direction of Edward Tufte.
posted by njohnson23 at 4:03 PM on December 21, 2021 [2 favorites]

Thanks! All the Covid modelling and tracking has me thinking a lot about data presentation, and I'm looking forward to digging into these resources a bit.

(There was a pair of particularly heinous graphs in the last government powerpoint I looked at.)
posted by invokeuse at 9:35 PM on December 21, 2021

I hadn't seen this! Nature Methods collections are often quite helpful.

One pet peeve: "Fancy" visualizations with high, multi-dimensional information density or relatively novel views of the data appearing in talks. Papers' visualizations may serve more purposes, but in a talk it is to clearly communicate something you've concluded about the data. As people get more familiar with software visualization options there's a tendency to come up with something cool that was helpful while exploring the data and then drop it in a slide that gives the audience no chance to digest it while the speaker talks about other stuff like methods for a minute or two.

Or, to put it in classic terms, don't get so excited about the visualization you can make that you forget to ask if you should make it.
posted by mark k at 11:39 AM on December 22, 2021 [1 favorite]