382
$\begingroup$

Data analysis cartoons can be useful for many reasons: they help communicate; they show that quantitative people have a sense of humor too; they can instigate good teaching moments; and they can help us remember important principles and lessons.

This is one of my favorites:

XKCD irony about correlation and causation

As a service to those who value this kind of resource, please share your favorite data analysis cartoon. They probably don't need any explanation (if they do, they're probably not good cartoons!) As always, one entry per answer. (This is in the vein of the Stack Overflow question What’s your favorite “programmer” cartoon?.)

P.S. Do not hotlink the cartoon without the site's permission please.

$\endgroup$
9
  • $\begingroup$ @sharpie: are jokes out? We obviously don't want the entire site to be humor, but everyone benefits from a little educational humor in small doses. $\endgroup$
    – Shane
    Jul 22, 2010 at 5:15
  • $\begingroup$ @Sharpie, feel free to close or reopen according to your feelings! I agree with Shane, a bit is ok, but not too much. For example, this question already included a funny cartoon. The jokes question not really a funny joke.... $\endgroup$
    – Peter Smit
    Jul 22, 2010 at 13:58
  • 43
    $\begingroup$ These cartoons are useful too; they can be included in a lecture on a particular topic where you are trying to explain a concept (e.g. correlation/causation above). A little humor can help to keep an audience engaged. $\endgroup$
    – Shane
    Jul 22, 2010 at 14:22
  • 9
    $\begingroup$ According to the tour, this question should be closed, since it is a question that has "too many possible answers" and since it is "primarily opinion-based". I'm not complaining, just surprised it has stayed open for this long. $\endgroup$
    – Flimm
    Dec 9, 2014 at 10:29
  • 5
    $\begingroup$ Data Science analogy to cartoon in OP. Data Scientist: I went to data science bootcamp and learned how to find correlations in big data. Those insights can be converted into big money. Statistician: But many of those correlations are spurious. Correlation does not imply causation. Data Scientist: Don't give me none of that century old statistics mumbo-jumbo. This is big data. That means the data has everything. So by definition, all relationships in the data are correct. I ring the cash register while you snooze and lose, grandpa. $\endgroup$ Dec 19, 2015 at 22:42

84 Answers 84

246
$\begingroup$

Was XKCD, so time for Dilbert:

alt text

Source: http://dilbert.com/strip/2001-10-25

$\endgroup$
3
  • 40
    $\begingroup$ Did anyone else notice that the tour guide changes colors between the second and third frames? $\endgroup$ Aug 22, 2012 at 20:41
  • 9
    $\begingroup$ On RANDU: "We guarantee that each number is random individually, but we don't guarantee that more than one of them is random." $\endgroup$ Mar 18, 2014 at 15:11
  • $\begingroup$ Link not working, was it this one dilbert.com/strip/2001-10-25 ? $\endgroup$
    – Tim
    Jan 27, 2015 at 19:27
229
$\begingroup$

Another from XKCD: ... okay, but because you said that, we're breaking up.

Mentioned here and here.

$\endgroup$
2
  • 12
    $\begingroup$ You can't read this one without the alt text. it said something like "But because of that we're totally breaking up" $\endgroup$ Mar 10, 2014 at 18:16
  • $\begingroup$ @generic_user The alt text (mouseover text) reads: "... okay, but because you said that, we're breaking up." $\endgroup$ Nov 8, 2020 at 1:05
211
$\begingroup$

My favourite Dilbert cartoon:

enter image description here

Source: http://dilbert.com/strip/2008-05-07

$\endgroup$
2
  • 11
    $\begingroup$ Definitively my favorite cartoon about Data Mining $\endgroup$
    – steffen
    Dec 1, 2010 at 12:24
  • $\begingroup$ Link not working. Was it this one: dilbert.com/strip/2008-05-07 ? $\endgroup$
    – Tim
    Jan 27, 2015 at 19:27
199
$\begingroup$

One more Dilbert cartoon:

http://dilbert.com/fast/2008-05-08/ ...

$\endgroup$
2
  • 8
    $\begingroup$ This one reminds me of the recent bailout in the States, where they just made up 700 billion number - they said they just wanted a really large number. :) $\endgroup$ Aug 12, 2010 at 8:53
  • 1
    $\begingroup$ Fixed. I had to add some dots after the cartoon since SE didn't allow me to submit the changes :-\ $\endgroup$ Feb 5, 2015 at 11:30
182
$\begingroup$

One of my favorites from xckd:

Random Number

RFC 1149.5 specifies 4 as the standard IEEE-vetted random number.

RFC 1149.5 specifies 4 as the standard IEEE-vetted random number.

$\endgroup$
1
  • 1
    $\begingroup$ But that isn't even prime! $\endgroup$ Jan 25, 2017 at 15:27
156
$\begingroup$

XKCD: significant

'So, uh, we did the green study again and got no link. It was probably a--' 'RESEARCH CONFLICTED ON GREEN JELLY BEAN/ACNE LINK; MORE STUDY RECOMMENDED!'

xkcd: significant

$\endgroup$
4
  • 19
    $\begingroup$ This is by far my favorite cartoon of all time. It's super educational. It really gets to the heart of the definition of a p-value. In fact, I bet that less than 10% the students who pass a college freshman "intro to stats" class get this joke, and this makes me sad. $\endgroup$ Jan 15, 2014 at 3:36
  • 5
    $\begingroup$ Maybe so! Fortunately for freshmen, @Glen_b has offered an excellent breakdown here. $\endgroup$ Feb 27, 2014 at 1:03
  • 2
    $\begingroup$ Great! But yellow appears twice :P $\endgroup$
    – Rodrigo
    Mar 1, 2016 at 0:51
  • 4
    $\begingroup$ this is a pretty good joke as it clearly demonstrates why repeated multiple testing is dangerous. For anyone interested check out Bonferi correction to deal with this. $\endgroup$
    – user46925
    Mar 7, 2016 at 18:21
156
$\begingroup$

Normal Versus Paranormal Distribution

From: A visual comparison of normal and paranormal distributions Matthew Freeman J Epidemiol Community Health 2006;60:6. Lower caption says 'Paranormal Distribution' - no idea why the graphical artifact is occuring.

$\endgroup$
5
  • 3
    $\begingroup$ I think this version of the joke works better (from oneweirdkerneltrick.com), though apparently this version was seven years earlier. $\endgroup$
    – Danica
    Feb 3, 2015 at 1:06
  • $\begingroup$ this isn't really funny. it's more of a twist on english terms $\endgroup$
    – user46925
    Mar 7, 2016 at 18:20
  • 1
    $\begingroup$ @zero "A twist on English terms" describes a great many jokes $\endgroup$
    – Fomite
    Mar 7, 2016 at 18:22
  • $\begingroup$ Yeah - I think all those jokes suck. There is no underlying statistical humour. This joke should be put on the English stackexchange instead. $\endgroup$
    – user46925
    Mar 7, 2016 at 18:23
  • $\begingroup$ @Phil: In the 2-dimensional version linked by Dougal, the "paranormal distribution" is indeed a (bizarrely truncated) distribution; so the joke has some statistical content. It doesn't work in one dimension, where your comment certainly applies. $\endgroup$ Jun 19, 2016 at 17:47
144
$\begingroup$

I just came across this and loved it:

alt text

(http://xkcd.com/795/).

$\endgroup$
133
$\begingroup$

Image at bp1.blogger.com.

$\endgroup$
3
  • 9
    $\begingroup$ That's great. The standard way of dealing with outliers. $\endgroup$
    – Shane
    Jul 26, 2010 at 20:06
  • 2
    $\begingroup$ Who's the artist? $\endgroup$
    – blubb
    Sep 25, 2014 at 16:01
  • 1
    $\begingroup$ This cartoon was drawn by Ben Shabad $\endgroup$
    – ff524
    Jan 7, 2016 at 20:20
125
$\begingroup$

Another from xkcd #833:

And if you labeled your axes, I could tell you exactly how MUCH better.

And if you labeled your axes, I could tell you exactly how MUCH better.

$\endgroup$
118
$\begingroup$

http://imgs.xkcd.com/comics/extrapolating.png

By the third trimester, there will be hundreds of babies inside you.

Also from XKCD

$\endgroup$
111
$\begingroup$

This isn't technically a cartoon, but close enough:

$\endgroup$
3
  • 10
    $\begingroup$ That's definitely my favorite. I always have to stop on this and laugh when scrolling over this page. It's just so bad!! $\endgroup$
    – Henrik
    Sep 9, 2010 at 9:02
  • 2
    $\begingroup$ +1 'Natch, Randal Munroe has a sweet variation on this. $\endgroup$
    – Alexis
    Sep 24, 2019 at 15:29
  • $\begingroup$ Yep: a pie is a pie chart telling you how much pie is left. $\endgroup$ Mar 31, 2020 at 21:03
106
$\begingroup$

Nice. The importance of variance when thinking about a population.

enter image description here

Saturday Morning Breakfast Cereal

$\endgroup$
0
103
$\begingroup$

this too: alt text

$\endgroup$
3
  • 4
    $\begingroup$ To be honest, those are the bad physicists. The good physicists stick around and make a name for themselves. $\endgroup$
    – Iterator
    Aug 6, 2011 at 3:17
  • 3
    $\begingroup$ Yet it's amazing how often it works... $\endgroup$
    – wnoise
    Sep 22, 2011 at 4:44
  • $\begingroup$ Economics defined $\endgroup$
    – luchonacho
    Oct 18, 2021 at 19:40
99
$\begingroup$

There is this one on Bayesian learning:

alt text

$\endgroup$
2
  • 1
    $\begingroup$ What's the source? $\endgroup$
    – Shane
    Aug 31, 2010 at 15:02
  • 1
    $\begingroup$ It was taken from Mike West's website: stat.duke.edu/~mw/fineart.html $\endgroup$
    – ebony1
    Aug 31, 2010 at 16:23
94
$\begingroup$

And another one from xkcd.

Title: Self-Description

alt text

The mouseover text:

The contents of any one panel are dependent on the contents of every panel including itself. The graph of panel dependencies is complete and bidirectional, and each node has a loop. The mouseover text has two hundred and forty-two characters.

$\endgroup$
0
80
$\begingroup$

Here is a nice one (the inadequacy about average ratings)
http://xkcd.com/937/

$\endgroup$
1
  • $\begingroup$ Kinda disappointed that Munroe did not work "selection bias" into the text. $\endgroup$
    – Alexis
    Sep 24, 2019 at 15:32
78
$\begingroup$

Another one from xkcd:

Hell, my eighth grade science class managed to conclusively reject it just based on a classroom experiment. It's pretty sad to hear about million-dollar research teams who can't even manage that.

Alt-text:

Hell, my eighth grade science class managed to conclusively reject it just based on a classroom experiment. It's pretty sad to hear about million-dollar research teams who can't even manage that.

$\endgroup$
0
74
$\begingroup$

Here's another one from Dilbert:

enter image description here

$\endgroup$
1
  • $\begingroup$ Looks like this one needs an updated image. $\endgroup$ Feb 24, 2015 at 20:45
66
$\begingroup$

More about design and power than analysis, but I like this one

alt text

$\endgroup$
66
$\begingroup$

I liked this one:

enter image description here

This is probably fun to show in class as well...

$\endgroup$
57
$\begingroup$

A classic...

Never discuss multiple testing without it

$\endgroup$
4
  • $\begingroup$ "Because medical research findings can be difficult to reconcile, are not always pre-digested, and can seem overwhelming to us casual observers, let us make fun of those who dedicate their lives to obtaining them." $\endgroup$
    – rolando2
    Feb 9, 2014 at 0:53
  • 16
    $\begingroup$ @rolando2 As a medical researcher, I find the sensationalist incompetence of mainstream science reporters hilarious. $\endgroup$
    – Superbest
    Feb 26, 2014 at 12:08
  • $\begingroup$ There's a listserv from HealthNewsReview devoted to evaluating media handling of health research findings. $\endgroup$
    – rolando2
    Mar 9, 2015 at 14:14
  • 5
    $\begingroup$ As a statistician, I find the sensationalist incompetence of mainstream "data analysts" hilarious. $\endgroup$ Mar 7, 2016 at 21:35
57
$\begingroup$

Source: unknown. Posted on flowingdata.com.

$\endgroup$
1
  • $\begingroup$ There could be a secondary y-axis labeled "Ballerina chicken with cloaca". $\endgroup$ Jan 31, 2019 at 9:38
56
$\begingroup$

enter image description here

Saturday Morning Breakfast Cereal

$\endgroup$
1
  • 3
    $\begingroup$ yeah but.... this one isn't true... it mostly depends on how you parameterize the time variable $t$... i guess if you go back far enough, but come on... $\endgroup$
    – William
    Sep 22, 2011 at 17:24
55
$\begingroup$

enter image description here

Found this one in the comments on Andrew Gelman's blog.

$\endgroup$
1
  • $\begingroup$ This chart is wrong, correct percentage is about fifty-fifty. $\endgroup$
    – juliohm
    Mar 10, 2014 at 16:09
54
$\begingroup$

I found this from a NoSQL presentation, but the cartoon can be found directly at

http://browsertoolkit.com/fault-tolerance.png

alt text

$\endgroup$
1
  • 12
    $\begingroup$ Can you please explain this cartoon? $\endgroup$
    – Christian
    Sep 3, 2014 at 0:00
48
$\begingroup$

Allright, I think this one is hilarious- but let's see if it passes the Statistical Analysis Miller test.

Fermirotica

I love how Google handles dimensional analysis.  Stats are ballpark and vary wildly by time of day and whether your mom is in town.

I love how Google handles dimensional analysis. Stats are ballpark and vary wildly by time of day and whether your mom is in town.

$\endgroup$
1
  • 19
    $\begingroup$ Statistical voyeurism? And there we were wondering what to call the site... $\endgroup$
    – walkytalky
    Jul 23, 2010 at 15:48
46
$\begingroup$

From xkcd:

enter image description here This is data analysis in the form of a cartoon, and I find it particularly poignant.

The universe is probably littered with the one-planet graves of cultures which made the sensible economic decision that there's no good reason to go into space--each discovered, studied, and remembered by the ones who made the irrational decision.

$\endgroup$
0
44
$\begingroup$

Another one from xkcd:

Coconuts are so far down to the left they couldn't be fit on the chart.  Ever spent half an hour trying to open a coconut with a rock?  Down with coconuts.

$\endgroup$
8
  • 14
    $\begingroup$ Bananas are always tasty. $\endgroup$ Dec 1, 2010 at 4:00
  • 2
    $\begingroup$ @AnonymousType and easy! $\endgroup$
    – tdc
    Feb 23, 2012 at 13:07
  • $\begingroup$ Where are persimmons? $\endgroup$ May 18, 2013 at 3:18
  • $\begingroup$ @EngrStudent: they're simultaneously off both ends of the tasty scale. $\endgroup$
    – naught101
    Aug 21, 2013 at 1:35
  • 4
    $\begingroup$ I'm surprised Durian isn't on here... $\endgroup$
    – Frank H.
    Oct 20, 2015 at 17:41

Not the answer you're looking for? Browse other questions tagged or ask your own question.