« Are new SEC rules enough to prevent another Flash Crash? | Main | Because it's Friday: Do my ears deceive me? »

September 23, 2011

TrackBack

TrackBack URL for this entry:
http://www.typepad.com/services/trackback/6a010534b1db25970b015391d21543970b

Listed below are links to weblogs that reference Data Visualization doesn't need to be biased:

Comments

Feed You can follow this conversation by subscribing to the comment feed for this post.

Thanks for the comments Dave!

I think the key point of the talk I'd like your readers to understand is that we have to start parsing our ideas, attitudes and best practices about data visualization between dataviz for analytical purposes and dataviz for communicative purposes. I think that analytical visualizations can and should strive for objectivity, but if my goal is to communicate a piece of information efficiently in a crowded information space, then we should use the tools of design to increase the visualization's effectiveness.

So long as communicative visualizations are done transparently and ethically, I see no reason not to encourage the use of visual flourishes.

Visualizations that purposefully obfuscate or are intentionally dishonest should be derided for what they are: propoganda. But a visualization adorned with meaningful and transparent visual flourishes is, in my opinion, okay. Rhetoric wrapped words have given us some of our most uplifting political speeches...why can't we do the same with our visualizations?

David and Alex,

I am sympathetic to both of your positions. I think some of the disagreement lies in what you want your graphics to do. Alex wants graphs to make a point in an authoritative fashion, such as how an expert tells non-experts how something works. David wants graphs to allows readers to see the interesting relationships for themselves. When I make graphs, I definitely follow David's approach.

I don't care if graphs push analyses on readers. Plotting relevant comparisons is itself pushing an analysis on the reader, although to a lesser degree than illustrated above.

My objection to the chart on the right above is that it pushes an analysis

    at the expense of clear comparisons
. This is a property of the chartjunk though, not of the intent of the creator. Whether chartjunk is intended to promote a conclusion or simply make a graph prettier, it is counterproductive to my goals when it gets in the way of clear comparisons. More here http://wp.me/p1RDVT-4S

Well, never forget: "figures don't lie, but liars figure". In the politically charged atmosphere of The Great Recession, being straightforward will be viewed as self-inflicted punishment. IOW, Social Darwinism takes the lead.

"Good charts ...present data, and leave the analysis ... to the viewer." Well said. Furthermore, there is a great advantage to the reader to see something that looks familiar. We, as analysts, should not force the reader to figure out some new graphic if a simple, familiar graphic suffices. To do so is to put the spotlight on ourselves ("Look at how clever I am!") rather than on the data.

Point of order. The image on the right is originally by Nigel Holmes, but the paper is about Nigel Holmes by authors at the University of Saskatewan.

Thanks Cody and others for pointing out that paper wasn't by Nigel Holmes. I've corrected the post above.

The comments to this entry are closed.


R for the Enterprise

Got comments or suggestions for the blog editor?
Email David Smith.
Follow revodavid on Twitter Follow David on Twitter: @revodavid

Search Revolutions Blog