Friends don't let friends make bad graphs

On the one hand, this all seems pretty great.<p>On the other hand, I think a lot of these &quot;bad graphs&quot; are very intentionally chosen precisely in order to hide the small number of data points, or an underlying distribution that looks suspicious, etc.<p>So it&#x27;s not so much &quot;friends don&#x27;t let friends&quot;, but more &quot;when you see a graph that chooses to obfuscate rather than clarify, suspect it might be intentional&quot;.

The example for &quot;Friends don&#x27;t let friends make heatmaps without maxing out outliers&quot;[1] is so common. It&#x27;s also very frequent in stats visualisations in videogames. If you play strategy or simulation games they often have visualisations to help the player understand what&#x27;s going on&#x2F;wrong, but for heatmaps because of the effect of outliers the heatmap gradient is often pretty useless. For example in the game oxygen not included, frequently if you do the temperature viz, everything just turns to either blue or some shade of pinky-red, because if you have a volcano or other heat source all the other colours seem cool. So you can&#x27;t distinguish between a 1000C volcano and your slightly overheating 270C steam room for instance, they&#x27;ll just be a pretty uniform shade of pinky red, and your overheating 60C base will be blue because it&#x27;s pretty cold relative to them. Meaning that heatmap is pretty much useless for diagnosing a bunch of temperature problems.<p>[1] Can&#x27;t remember the exact wording but that one it&#x27;s like number 6 or 7.

A while ago i had a heated discussion on HN with someone who claimed that any graph where 0 is not the minimum value of all the axes is misleading.<p>We were talking about a graph that shows global temperature rise due to climate change. They claimed the graph was misleading because the Y axis (temperature) didn&#x27;t start from 0 (fahrenheit? celsius? fucking kelvin?).<p>This person also quipped, &quot;maybe if you can&#x27;t see if with 0 at the bottom, it&#x27;s not such a significant change?&quot;. That put a dent in my faith in humanity for a while. I&#x27;m just glad to see us operating at a higher level. I guess 2016-2020 was a different time.

A great reference for further reading about data visualization is &quot;The Visual Display of Quantitative Information&quot; by Edward Tufte, a classic book originally published in 1983 with enduring relevance today.

This is a great overview of common mistakes in data viz I will be sharing it with my colleagues. As a good supplement I highly recommend Kennedy Eliot&#x27;s &quot;39 studies about human perception in 30 mins&quot; <a href="https:&#x2F;&#x2F;medium.com&#x2F;@kennelliott&#x2F;39-studies-about-human-perception-in-30-minutes-4728f9e31a73" rel="nofollow noreferrer">https:&#x2F;&#x2F;medium.com&#x2F;@kennelliott&#x2F;39-studies-about-human-perce...</a><p>... a whistle stop tour around the research basis for a lot of claims around data viz best practice (esp interesting re the dogma around not using pie charts which seems to be a consistent bugbear of designers going back to the 1930&#x27;s but around which the research is inconclusive at best)<p>[edit: fixed a typo]

<i>&gt; 3. Friends Don&#x27;t Let Friends Use Bidirectional Color Scales for Unidirectional Data</i><p>why use colors at all in those examples?

Quite a lot of these lessons are not new, see Willard C. Brinton &quot;Graphic presentation&quot; - a book (1939), freely accessible <a href="https:&#x2F;&#x2F;archive.org&#x2F;details&#x2F;graphicpresentat00brinrich&#x2F;mode&#x2F;thumb?view=theater" rel="nofollow noreferrer">https:&#x2F;&#x2F;archive.org&#x2F;details&#x2F;graphicpresentat00brinrich&#x2F;mode&#x2F;...</a>.

These are still all bad to my eye. Way too much &quot;chartjunk&quot;, too many colors for most of them. You simply don&#x27;t need all those lines. They are easier to read if you put less on them, but carefully. Any edward tufte book is about this and a handful of basic techniques go a long way.

I have never liked violin plots,but I am very much not in the field of data visualisation. Just a week ago I stumbled on a video titled &quot;violin plots should not exist&quot; which made things fall into place: <a href="https:&#x2F;&#x2F;youtu.be&#x2F;_0QMKFzW9fw?feature=shared" rel="nofollow noreferrer">https:&#x2F;&#x2F;youtu.be&#x2F;_0QMKFzW9fw?feature=shared</a>

I wrote a talk entitled How Humans See Data that puts several of these ideas, among others, into a coherent framework based on research by Bill Cleveland.<p><a href="https:&#x2F;&#x2F;www.youtube.com&#x2F;watch?v=fSgEeI2Xpdc">https:&#x2F;&#x2F;www.youtube.com&#x2F;watch?v=fSgEeI2Xpdc</a>

I remember reading a guide a while back, just a web page... from someone who just seemed like they were a master at picking the right graph&#x2F;chart style for different types of data. He approached it in an exhaustive way by breaking down data into categories and explaining, for each data style, exactly why each chart style and not another one should be picked. I really wish I had bookmarked that guide because it just made it seem so straightforward.

One other rule I&#x27;ve found helpful is banking to 45 degrees -- it&#x27;s an easy way to rmake elationships easier to perceive: <a href="http:&#x2F;&#x2F;vis.stanford.edu&#x2F;files&#x2F;2006-Banking-InfoVis.pdf" rel="nofollow noreferrer">http:&#x2F;&#x2F;vis.stanford.edu&#x2F;files&#x2F;2006-Banking-InfoVis.pdf</a>

Never understood the hate for pie charts when stacked bar charts have a similar problem (hard to compare groups when they&#x27;re in different positions). If a pie chart isn&#x27;t good enough, probably switch to a paired bar chart or some other pair-wise comparison.

I still generally attempt to follow the rules established by Edward Tufte in The Visual Display of Quantitative Information.<p>Basically, Tufte used the idea of &quot;ink&quot;, classified into two groups, data ink and useless ink. The goal is to have a graph with as little useless ink as possible; where every bit of information visible in a graph (or table) is relevant to the end output. To this extent, he recommended dumping axis lines where unneeded, labels, keys, gridlines, and many more things.<p>LaTeX tables, by default, tend to look like what Tufte proposed, which is probably why LaTeX tables look so damn good compared to the HTML defaults

Writing an &quot;opinionated essay&quot; in a github readme is the most 2023 dev thing I&#x27;ve seen

Nice.<p>But I&#x27;d add getting rid of heatmaps on large datasets. They are information dense and pretty, but I can&#x27;t see how anyone interprets them. Better to do clustering and plot the data for each relevant cluster in a more meaningful way.<p>Just a thought

Related - Kaiser Fung has a long-running blog on bad graphs:<p><a href="https:&#x2F;&#x2F;junkcharts.typepad.com&#x2F;" rel="nofollow noreferrer">https:&#x2F;&#x2F;junkcharts.typepad.com&#x2F;</a>

Also: friends don&#x27;t let friends use line graphs when the data is not continous between the sampled points.<p>(bar graphs are typically more appropriate)

A more basic page might serve as a valuable introduction for some.<p>e.g. How many graphs have you seen on evening news programs that don&#x27;t even have axes labels?

Ironically, I was about ten examples in before I even noticed the tiny good&#x2F;bad labels.

I wish friends would call them charts and not graphs.