Also a good meta example of picking a poor visualization method for the given data?
This makes me curious --
Has there been any good application of "periodic table" as a visualization method for anything other than chemical elements?
If we are thinking of a grid where "rows" and "columns" imply some common properties .. then I am sure there are a few examples of matrix / quadrant/ grid formats that help convey that.
But specifically "periodic table" is overused IMO.
If applied correctly, I would expect every Nth element along X dimension resembles each other so it makes sense to wrap it and show it as a next row so that we get common properties along Y dimension as well.
I think it's an amusing display of a bunch of different visualization methods.
It's fun to browse. No need to take it more seriously than it takes itself.
From their white paper conclusion:
---
Our efforts in structuring the vast domain of visualization
methods cannot be seen as a close adaptation of the periodic table of chemical elements. It is rather a functional,
metaphoric homage to it. The choice of methods included
as well as the order criteria cannot be considered exhaustive. Nevertheless, it does provide an overview over more
than hundred useful visualization methods of great variety
and by organizing them assists researchers and practitioners alike in choosing adequate visualization methods for
their needs. On demand the user is provided with further
useful information through signs. We encourage the
reader to playfully explore the different properties of the
visualization methods presented. So he may consider
more than one method for his next visualization requirement and to use them in a combined, complementary
manner. This may not turn lead into gold, but turn complex issues into accessible explanations.
---
I do accept that a not insignificant segment of HN users would much prefer a man-page-like display, all-text, no graphics, no css, no javascript-- even for a page about visualizations!
"Concept" and "Information" should have more distinguishable colors though.
> I do accept that a not insignificant segment of HN users would much prefer a man-page-like display, all-text, no graphics, no css, no javascript-- even for a page about visualizations!
I like your glass-half-full positive take, but I’m not sure this is accurate; most posts on anything graphical that link to a github page have lots and lots of complaints about the lack of screenshots or examples. People overwhelmingly like pictures, especially on pages about viz, and parent wasn’t at all suggesting or implying not to use a picture, but maybe implicitly to use a better & more relevant picture. The question here is whether the periodic table metaphor offers anything of value to a collection of viz examples, or if it actually confuses the viewer by suggesting or implying some kind of meaning or organization when there is none. There’s not a small amount of irony in using a bad metaphor for visualizing something in an article that is specifically discussing how to use good metaphors for visualization, right?
The problem with a lot of things people try to mush into something that looks like the periodic table is that most datasets:
• Have too few items
• Have too many items
• Simply don't fit into a near table, needing more dimensions, or having two obvious dimensions that line up into a grid less well than the elements do.
Heck, even the elements don't really fit. It just happens that it works & looks OK if you break out a load from the bottom left to stop it being too wide. As someone else has already said: the periodic table is more a document of a model, that by chance happens to work well enough as a visualisation.
>Has there been any good application of "periodic table" as a visualization method for anything other than chemical elements?
No, because as can be seen in all of them the structure of the table maps to the structure of the chemical elements.
The visualization of chemical elements is successful because it is an accurate representation of how chemical elements work, it does not just represent data but also is mapped to a model of how that data can conceivably change in the future.
Every periodic table of X that is not chemical elements is stupid and does not work because
1. come on - what are the odds that some other subject in the universe has a similar model to chemical elements.
2. the visualization is not done to best represent the data but the data is bent as best as it can be to fit into the periodic table structure.
And what is meant by periodic structure here - it is meant a selection of cells laid out in a very specific way with a color scheme so that if you saw it far enough away you would instantly think - hey, it's a periodic table!
The periodic table used this way is not a data visualization method, it is a visualization parody
method.
on edit: also pisses me off because in my world a parody should at least be funny, darn it.
sorry for the rant, this stuff really pisses me off.
for example one thing I might want to have to describe visualization methods would be what graphical elements would be important to it.
In the case of the periodic table an important graphical element would be color (the color used carries date), an unimportant one would be typography.
this is obviously just thinking aloud here.
perhaps also some visualization methods would benefit by reference to their creators. sparklines for example. So it seems to me a graph of some sort would be useful.
well I think there's a natural hierarchy that emerges from the examples given in the periodic table link.
Spatial: line graphs, scatterplots, and even barplots and pie charts(1-dimensional space). Id throw spectogram and anything with time in there as well.
Geographic: maps (this could be a subset of spatial)
Graphical: graphs, mind-maps, flow charts, all the other graphs-like charts from the examples
Visual Metaphors: those weird examples of they give of steps and a bridge, a slide, etc
Those 4 categories alone capture all the examples from the webpage, and there are plenty of subcategories within them.
So to me a hierarchy / tree seems like a good choice, which is essentially a directed graph (like you suggested)
they're groups so symmetries are kinda what they do. There are families etc. in columns, with the final column being the generator. Then the sporadic groups are the rare earth elements which don't define a family of groups. order is increasing on the way down, and the very obvious cyclic groups are the noble gases equivalent.
The International Phonetic Alphabet might be a good contender. There’s no periodicity, but the columns for the consonants roughly represent place of articulation. The spacial arrangement of vowels is even closer to where in the mouth the sound is formed.
When I was learning to read Japanese the Hiragana alphabet was displayed in a somewhat periodic table-esque format. The vowel ending sounds were the vertical axis (a,i,u,e,o) then next column was ka, ki, ku, ke, ko - then sa, shi, su, se, so etc.
I like the idea, but I'm confused. Why go to all the effort to think about visualization methods, then shoehorn them into the shape of the chemical element periodic table, without even making use of the periodic structure of the table? Funny enough, "periodic table" isn't its own item on this chart.
I think it would be more useful and interesting to group by type in some way, perhaps with a bit of tree structure. Which reminds me of a loosely related idea: I'd like to see a tree diagram of different tree data structures, where the tree structure reflects the relationships of historical development and/or data structure behavior...
This is interesting and a little bit entertaining.
There has been excellent work on taxonomies for data visualization. Tamara Munzner's book "Visualization Analysis and Design" is a good starting point.
There has been excellent work on identifying "dimensions" or properties of data visualization. Lee Wilkinson's "Grammar of Graphics" is a landmark. You could search backward to find Bertin's Semiology of Graphics, or forward to find ongoing work like UW's Vega https://idl.cs.washington.edu/papers/vega-lite/
Some of this work might somewhat capture the properties of a periodic table to organize a catalog and even support further discovery.
I ironically love this and think it's (un)intentially hilarious. It's basically everything people who think very seriously about infoviz warn about, crammed into a single terrible visualization on the subject of visualization. It's such an erudite accumulation of inexpert concepts it almost has to have been put together (paradoxically) by an expert in the field.
I looked it up, and indeed it was put together by a couple serious folks [1]. Even extensively cites Tufte.
I reminds me of the movie "Bad CGI Gator" from last year.
This is so wonderfully hard to parse as either parody or serious. Brilliant.
Wait, is it parody? I would agree if it was, but I assumed it was serious, and kinda bad. Their “knowledge maps” page on the site looks similar - the examples are ugly & awful. https://www.visual-literacy.org/maps/ This all looks like it’s intended to be serious, considered good by the authors, especially considering the associated academic publication, and not some kind of intentionally funny commentary on bad visualizations. If that’s the case, it doesn’t seem particularly brilliant.
Aside from the nonsensical analogy to the periodic table (which they admit in the paper), the thing that bugs me about it is that it’s nowhere near complete, it does not cover all basic elements of visualization (also which they are aware of and discuss in the paper), so just completely misses the fundamental idea. It amounts to a hundred random examples of info-viz, most of which look mediocre to bad.
oh my. Maybe it is? The university they're attached to seems to be a real, serious, school. But these examples are just...worst of the worst stuff. Inscrutable metaphors, wasted axes, variable fonts without meaning, organization schemes that make no sense...the tribute to Ben Shneiderman literally breaks most of his interface design rules and his goals and theories on infoviz. [1]
I understand that much of this material is relatively old, but we actually were better back then.
The "tell" to me (if this is parody) is the percent of the audience here who can't tell if this is serious, or good, or not.
Perhaps somebody should count the comments and put them in a pie chart inside of a pie store and call it "HN taste tests visual-literacy.org".
Okay, I'm calling it, this is a very inside joke page of some sort.
I don't know if I am missing something obvious. What is the periodicity in these visualization methods? Like I am looking at the 2nd column: cartesian coordinates, line chart, area chart, scatterplot, spectogram. What periodic property is being repeated among them?
There's no periodicity in visualization. The design space of visualization is much more complex the atomic number.
Also, this "periodic table" does not show the three main branches of visualization: Information Visualization, Scientific Visualization, and Geographic Visualization.
For example, this page gives an overview of time-oriented visualization with meta information like the original paper: https://browser.timeviz.net/
It also seems to be maintained by people working in that area.
Wow, this is terrible. It's an inappropriate metaphor, it's badly executed, the examples for each category are hardly representative and often spectacularly bad (bar chart!?), and to call it "incomplete" would be a severe understatement.
This is a great example of how not to organize and display information.
That's an analogy I can work with. Lots of those visualisation methods are not fit for human consumption. And in fact a whole lot of them are just grey blobs that shouldn't be near humans at all.
The point of the periodic table is that it supplies important structural information about each object in the set.
Using the shape of the periodic table rather than its philosophy when the stated intent is to teach about appropriate data representation is an immediate wtf for me.
Great idea, but maybe it would be easier to use if you were using tags or something similar. You could show all the images at once and allow user to filter the view.
I'd like to see the practice of stuffing arbitrary topics into the periodic table become an object of scorn and eyerolls. Doing so degrades both the topic and the science of chemistry.
Why is the "noble" (rightmost) column filled with garbage visualization methods? Seems like very very little thought went into the "period table" aspect of this
I feel like all these "periodic table of X" things completely miss the point of why the periodic table of elements is the way it is, and why it's a fantastic tool to organize elements and nothing else.
1) It's not clear for me how color segmentation groups were picked in the first place. Are they mutually exclusive?
2) Would be great to make this interactive. There is a substantial cognitive effort to imagine these methods from the name of few words (good exercise but come on).
This makes me curious --
Has there been any good application of "periodic table" as a visualization method for anything other than chemical elements?
If we are thinking of a grid where "rows" and "columns" imply some common properties .. then I am sure there are a few examples of matrix / quadrant/ grid formats that help convey that.
But specifically "periodic table" is overused IMO.
If applied correctly, I would expect every Nth element along X dimension resembles each other so it makes sense to wrap it and show it as a next row so that we get common properties along Y dimension as well.