Towards a Theory of Bullshit Visualization [pdf] | Hacker News Reader

Towards a Theory of Bullshit Visualization [pdf]

research.tableau.com

93 pointspolm2314 comments

"In this unhinged rant, I lay out my suspicion that a lot of visualiza- tions are bullshit: charts that do not have even the common decency to intentionally lie but are totally unconcerned about the state of the world or any practical utility. I suspect that bullshit charts take up a large fraction of the time and attention of actual visualization pro- ducers and consumers, and yet are seemingly absent from academic research into visualization design. ... It is my contention that many (most?) visualizations are bullshit in either the Frankfurt sense of attempting to persuade while being totally disconnected from notions of truths about the world, or in the Graeber sense of being created for no readily apparent purpose and to no real useful end. When they inform, these bullshit charts do so in the most superficial and unsatisfying ways— for instance, a bar chart might tell me that so-and-so many widgets were sold during a particular period of time"

> "for instance, a bar chart might tell me that so-and-so many widgets were sold during a particular period of time"

Funny, considering that's a core part of Tableau's value proposition.

If their product becomes saturated with bullshit charts it will hurt them in the long run.

Also, it's a sign that at the very least this person working there genuinely cares about the product they're selling, so that's good

So disappointed :) I was hoping this was going to be a visualization OF bullshit, not a rant about bullshit visualizations.

All seriousness aside, Correll fails to realize that the layout, length, and organization of your text are also forms of visualization, and at those, he fails completely.

"It is my contention that many (most?) visualizations are bullshit in either the Frankfurt sense of attempting to persuade while being totally disconnected from notions of truths about the world, or in the Graeber sense of being created for no readily apparent purpose and to no real useful end."

So he's shown us some erudition by throwing in Frankfurt & Graeber. Whoopee. When he mentions Tufte, a pioneer and a giant in his field, he's dismissive.

All that said, there is some worthwhile stuff in here. He needs to study The Elements of Style, hire an editor, and reduce it to a readable three pages or so, single-column.

Have you ever read an academic paper? This is a pretty bog standard layout.

Indeed I have.

Academic papers -- the gold standard for readability!

Said no one, ever.

A more sophisticated exploration of this topic is Calling Bullshit[1], by Carl Bergstrom and Jevin West. That's a book, but the linked website also has a lot of freely available resources. It's been taught as a college course - week 6 is data visualization.

[1]: https://www.callingbullshit.org/

I feel this would have been more suitable as a blog post rather than a two-column layout under a "research" subdomain. I wonder if "Tableau Research" has a rigorous peer-review process for its research publications.

The layout and URL are part of the joke. The article opens with "In this unhinged rant"--I really don't think this is any sort of attempt to fool the reader.

With respect, this author misunderstands Edward Tufte's use of the term "Chart junk." Chart junk isn't just pointless decoration on charts. It's also misleading or missing scales, deceptive or inept ways of representing data, and so forth.

Of course Tufte's body of work a whole, and VDQI in particular address those issues, but chartjunk is, in my reading, a specific and limited complaint. His leading two examples are unintended moiré effect in cross hatching unnecessary grid lines. His third (the duck) is a bit more nebulous, but I read it as any sort of overly cute “we can so we did” effect on a visualization.

The mortal sins of data visualization misrepresent the data; the venal sin of chartjunk distracts from the data.

Talking about bullshit visualizations I think this should top the chart:

https://cms.qz.com/wp-content/uploads/2018/09/Modi_Chart.jpg...

https://qz.com/india/1388030/indias-bjp-peddled-a-prime-exam...

No, this is not what's being talked about. That's a deceptive chart, not their idea of Bullshit Visualization's. It's an example of the minority of bad charts, as your link shows it gets jumped on quick.

This from your article is a better example -

https://theatlas.com/charts/HkBllCLO7

What does that mean? Is it CPI? Is it consistent with worlds prices? Because of inflation prices go up over time. It goes up over time. Is the chart saying anything?

Here's a good example of Bullshit Visualization's - https://xkcd.com/1138/

It's things you'll see in the media/PR Journal articles that are nothing but look pretty and give a story a feel of authority when it shouldn't.

Thanks for sharing this, gave me a smile and a good chuckle.

Some other people who are knowledgeable in the domain have posted that there might be some rigor issues.

But I think a lot of us can empathize with the feeling of "No you idiot, that's not how you use that at all!" when it comes to some area we know more about than the average person.

I'm pretty sure my car mechanic winces every time I bring my vehicle back into the shop. That's what I pay him for I guess.

Reply on news.ycombinator.com