click on the thumbnails to view the original PDF..

These are visualizations I did in response to the call at Visualize this for visualizations of Ken Wark’s Gamer Theory.

frackman.org/blog

This is a visualization of the only the nouns and verbs in the first chapter of the book (AGONY on The Cave(TM)), organized only according to the sentences and paragraphs. My main intent with this series was to try and combine all information available into a single visualization. So, I thought to use paragraphs, sentences, word frequency, part of speech tagging and anything else I could think of. I found it difficult to map too many dimensions however, and since I was doing graph based visualizations I wanted to keep the node count down, or Graphviz takes an eternity*.

I think this severely truncated text is somewhat understandable still. As an aside, evidently it is not as easy to read scrambled intra-word text as I thought.. I struggled with including “determiners” (here is the list of parts of speech that the OpenNLP part of speech tagger tags things with: Penn Treebank Parts of Speech) – it seemed egregious to me to eliminate the word no which is either classified as “DT” or “UH” – interjection, however there are other interjections which are less important. In the end, my rationalization for eliminating the concept of negation is that you can still (mentally) mine the underlying connections, and the logical comparison is less important than the connection. At least, thats my story and I’m sticking to it.. It has some strange consequences – “No pain no gain” in the text becomes pain -> gain in the sentence visualization. Red circles are verbs (any type, just looking for the “VB” prefix from the pos tags) and blue squares are nouns (also just looking for the “NN” prefix)..

For these two visualizations, I used the text of another chapter (ANALOG on Katamari Damacy) and knocked down to only the nouns. However, I wanted to actually use the graph structure to try and convey some information – the earlier paragraph/sentence visualization doesn’t have any edges – so I created an index of nouns, and then edges from the index word (in a black box) to any occurences in the text. I then rendered it using “dot” for the hierarchical one and “neato” for the energy minimized one. For the energy minimized one (the one I was really shooting for..) the nodes overlap, and I attempted to fix this using “neato’s” voronoi and scaling overlap features – sadly, the voronoi code seems to take forever.. I also tried the fix here, to no avail. The scaling just crashes Graphviz.. So, not what I wanted, but better than nothing.

	dot renders things more quickly than neato, I was impatient, and once I saw it I really liked it.
	this is closest to my original intention with the Katamari Damacy visualizations, since it resembles a big ball..

* I’d like to thank the authors of Graphviz for providing such a useful and generally robust tool. I’ve been looking at a lot of graph tools and libraries lately, and it generally produces the nicest looking graphs, despite any grumbling on my part. I’m usually more critical of things I tend to use alot, so take it as an endorsement! There is also a MacOS X port of Graphviz which is what I’ve been using.

A TextArc is a visual represention of a text—the entire text (twice!) on a single page. A combination of an index, concordance, and summary; it uses the viewer’s eye to help uncover meaning. Presented here are the full interactive Java application and a PDF download of a “raw draft” of the Gamer Theory TextArc. In the PDF, the strand clusters beneath words point to points in the arc where those words occur. In the Web version, simply clicking a word reveals its distribution along the arc.

Play with me

get pdf

Try out TextArc for yourself. Click on a link below to open TextArc and start visualizing. If you have questions, the interface provides a built in Help section. Please be patient while it loads—it has to do a lot of thinking to look as good as it does.

• Open TextArc in a window (recommended)
• Open TextArc full screen

TextArc currently works best with Firefox or IE7 on PCs and Macs that have the latest version of Java. Due to a known bug in the Java VM the browser’s cache of Java classes is not cleared between executions; this means to restart TextArc you must exit and restart the browser.

Also see: textarc.org

McKenzie Wark on TextArc as a writing tool

Earlier this April (2007), Brad Paley and McKenzie Wark spent an afternoon at the Institute for the Future of the Book exploring a large image of the Gamer Theory TextArc projected onto one of the office’s walls. Wark later reflected on the experience:

As a writer I’ve long been fascinated by the possibilities of text visualization. Are there ways in which picturing the text, or more precisely relations within the text, might become part of the writing process? This is a bit different to the more ‘scientific’ applications of visualization. I’m interested in those too, but for me its mainly about writing.

Most writers who have wrestled with large texts realize at some point that for every sentence on the page (or screen) there are thousands of other sentences that could be in its place. Could there be ways of exploring this ‘virtual’ space of the text? Perhaps as a way of generating new texts out of the existing one? This was the line of thought that led me to Gamer Theory 3.0: The Visualizations.

By far the most promising visualization project I have seen is Brad Paley’s Text Arc. It is at once well designed, intuitive to use, and it produces a rich picture of relationships within the text. I was very excited to meet Brad at the Institute for the Future of the Book and learn more about his work, but I was really itching to see Gamer Theory as a Text Arc.

Brad put it up on the screen and it was like seeing a diagram of my own writing brain. It’s a way of seeing the text that shows me what is where. The results are sometimes surprising. For example, the term ‘agon’ ends up more or less at 9 o’clock, when I thought it would cluster more at 1 o’clock. It is, after all, the nominal subject of the first chapter. What I had not realized is that it is only announced there, and is actually developed much later. Had I known this in advance I would have changed the text. Oh well. Next time I hope I can use Text Arc mid way through the writing process, not at the end.

I was also surprised by things that clustered together: ‘theory’ is close to ‘entertainment’, ‘life’ with ‘world’ and ‘topology’ with ‘space’. The last of these three is to be expected, but the others are not. So for me what would be interesting would be to tease out the meaning of these spatial overlaps, and see if there are unexplored conceptual overlaps. That is one way I would use it in the writing process. The term ‘lifeworld’ has a particular theoretical meaning. But I did not use the term in the book. But maybe the concept crept in anyway, without my really wanting it to. And, of course, there’s something uncanny about the ‘theory’ ‘entertainment’ overlap. I have thought about writing something about this, and maybe the thought is already in the text before I have gotten around to writing it. I would want to explore that further, but for me its a potentially stunning result: that one could find inside the actual text these ‘virtual’ texts (I use the term virtual here in the philosophical sense).

I’m also surprised by things that don’t cluster: ‘work’ and ‘play’ is an obvious one. And I would want to know more about why. I was trying to get away from a binary relation between the terms. Indeed the last chapter ends with a four part schema: worker, player, hacker, gamer. So maybe in this case I succeeded in repositioning the terms a little bit. Changing the scale a bit and looking at the rarer terms, I would want to explore the relationship between ‘actual’ and ‘form’ some more. I would also want to know why form is not a more common term. The relations between terms that appear at different scales in the text is also intriguing to me. It’s easy to do a word count and find how common a term is, but adding the ‘spatial’ aspect, of where it falls in the text, is much more helpful.

There’s more, but perhaps now it’s clear where I’m going with this. Text Arc could be a really interesting way of reading, but also of writing. One thing I wish it did was allow the reader to pull a relationship out of the circle and explore it as a ‘thread’. That would allow the reader (and writer) to ‘unspool’ the hairball a strand at a time, seeing where the strands intersect.

Another thing I would like to do with it is look for sentence forms, and chart where they are. The key terms are one thing, but the structure of clauses is quite another thing to map. For example, my last book used chiasmus a lot, but it would be good to know the frequency of the construction and its placement in relation to other constructions.

In short: Text Arc is more than a pretty picture. I hesitate to say that it is a ‘useful tool’, really because I think it is (and can be something more interesting than that. I see it more as a creative process than as an instrumental one. The ‘result’ it would give you is more of the order of the new writing you can already find in your existing writing. Walter Benjamin once said that “the work is the death mask of its conception.” I think with Text Arc I get to peel off the death mask of the finished text and find some life teeming away inside it, making new texts, that I might not otherwise know were there.

McKenzie Wark
18th April 2007

About the Artist

W. Bradford Paley uses computers to create visual displays with the goal of making readable, clear, and engaging expressions of complex data. He did his first computer graphics in 1973, founded Digital Image Design Incorporated (didi.com/brad) in 1982, and started doing financial & statistical data visualization in 1986. He has exhibited at the Museum of Modern Art; he is in the ARTPORT collection of the Whitney Museum of American Art; has received multiple grants and awards for both art and design, and his designs are at work every day in the hands of brokers on the floor of the New York Stock Exchange. He is an adjunct associate professor at Columbia University, and is director of Information Esthetics: a fledgling interdisciplinary group exploring the creation and interpretation of data representations that are both readable and esthetically satisfying.

We present ‘FeatureLens,’ an application dedicated to the visualization of collections of text documents based on text features. Our goal is to integrate the results of text-mining algorithms into a meaningful representation of a text collection. We choose to select parts of text according to the distribution of some features such as the word frequency across and within the chapters of a book. Trends, gaps, patterns and outliers in the distributions are used to select ‘interesting’ patterns in the documents.

Researchers

Anthony Don, HCIL
Machon Gregory
Elena Zheleva
Sureyya Tarkan
Catherine Plaisant, HCIL
Ben Shneiderman, HCIL
In collaboration with Tanya Clement (UMd Dept of English) and Loretta Auvil (NCSA)

FeatureLens Live!

Try FeatureLens live

Start by ‘loading’ Gamer Theory using the [LOAD] button in the upper left of the application.

Text Mining

We extracted one set of words and one set of frequent patterns of trigrams from ‘Gamer Theory 2.0′ using the Text-to-Knowledge framework (http://alg.ncsa.uiuc.edu/do/tools/t2k).

Stop words such as “a,” “the,” and “of” were filtered out from the set of words and the frequency (number of occurences) of each word was computed for each paragraph of the book.

A trigram is a set of three consecutive words. For example, in the text “the quick brown fox,” there are two trigrams: “the quick brown” and “quick brown fox”.

The text was preprocessed to support detection of trigrams that frequently co-occur in the same paragraphs. In our setting, a set of trigrams that occurs in at least three paragraphs is considered to be a frequent pattern of trigrams. Stop words were not filtered out for this analysis.
Frequent patterns of trigrams may appear when a sentence, composed by more than three words, is exactly repeated in different paragraphs or when some slight variations occur in the repeated sentence. The frequency of each frequent pattern was computed for each paragraph.

We will use the term “pattern” to refer to either words or frequent patterns of trigrams.

FeatureLens

Figure 1: The FeatureLens GUI

Figure 1 shows the Graphical Interface of FeatureLens. The interface is divided vertically into three parts.
In the leftmost part, the “Frequent Pattern” section contains a selection of words and frequent patterns. Patterns can be filtered by minimum size (number of trigrams in a pattern) and minimum frequency within the whole book. A text query can also be used to load patterns.
In the middle of the screen, the “Collection Overview” displays:

• the distribution of the selected patterns across the chapters of the book
• an overview of the chapters, using one gray line per paragraph
• the legend with the selected patterns and their associated color

In the rightmost part, the “Document View” displays the currently selected paragraph and its context on a blue background. The text is colored to show the position of the selected patterns inside a paragraph.

Sorting the distributions of pattern frequencies by trends

We use the distribution of individual pattern frequencies in each chapter to build different orderings for patterns. The user can sort patterns that:

• remain constant : retrieves patterns with the same frequency in each chapter. Patterns are ordered by decreasing total frequency.
• increase or decrease : retrieves patterns whose frequency steadily increases or decreases along the book. These patterns may represent topics that get more and more (resp. less and less) emphasis along the book.
• contain sink or spike : retrieves patterns that are used less (resp. more) in one chapter. These patterns may characterize a particular idea that is developed in one chapter.
• drop or float : retrieves patterns that have a high then low frequency or the other way around.
• contain gaps : retrieves patterns that start and finish with a high frequency but have a low frequency in the middle of the book.

Patterns can also be sorted according to their trend within a chapter. We provide four preselected trends for:

• patterns that have a low frequency in the beginning and at the end of the chapter but a high frequency in the middle .
• patterns that have a low frequency in the beginning and in the middle of the chapter but a high frequency at the end .
• patterns that have a high frequency in the beginning and at the end of the chapter but a low frequency in the middle .
• patterns that have a low frequency in the middle and at the end of the chapter but a high frequency in the beginning .

FeatureLens Examples

Five examples of FeatureLens that demonstrate how an analysis can lead to insights into the structure of the text.

Frequent patterns form “a line of a certain type”

Figure 2: Long Patterns

Figure 2: The patterns are sorted by increasing trend. The word “problem” is selected and the graph shows a steadily increasing frequency starting from the “Atopia” chapter. Is this trend meaningful? Does the author want to ask new questions and underline new problems as the book progresses? New hypotheses are being provoked and can be verified by reading the corresponding paragraphs.
Video (avi, no sound)

The rise of problems

Figure 3: The Rise of Problems

The patterns are sorted by increasing trend. The word “problem” is selected and the graph shows a steadily increasing frequency starting from the “Atopia” chapter. Is this trend meaningful? Does the author want to ask new questions and underline new problems as the book progresses? New hypotheses are being provoked and can be verified by reading the corresponding paragraphs.
Video (avi, no sound)

The real world disappears

Figure 4: The real world disappears

The patterns are sorted by decreasing trend. Two words appear on the top of the list “real” and “world.” The graph of these patterns shows a decreasing trend along chapters for the word “real.” By reading the corresponding paragraphs, it appears that “real” mainly appears in “real world.” Does this fading-out of the real world follow the author’s ideas about the transition from analog to digital world?
Video (avi, no sound)

The steep hill

Figure 5: The steep hill

The patterns are sorted by “spikyness.” On the top of the list, four patterns form a spike in the “Analog” chapter. These words are “sisyphus”, “prince”, “katamari” and “ball,” and they only occur in this chapter. Why is this chapter so particular? What is the topic? By reading the content of the chapter, it appears that it contains an analysis of the game “katamari damacy,” a game where the player controls a ball through various levels. It also contains a parallel between “the Myth of Sisyphus” and the neverending nature of the game.
Video (avi, no sound)

From “gamers” to “gamer as theorist”

Figure 6: From “gamers” to “gamer as theorist”

Figure 6: The patterns are sorted by trends inside the chapter entitled “Agony.” First, the patterns with a high frequency at the beginning are retrieved: the words “gamers” and “screen” have a high value. Then, the patterns with a high frequency at the end of the chapter are retrieved: the trigram “gamer as theorist” and “utopia” have a high value. It seems that in this section, the author switches between two concepts, “gamer” and “gamer as theorist.”
Video (avi, no sound)

Conclusion

The selection of trends in the distribution of pattern frequencies allows collecting meaningful pieces of information about the text. FeatureLens is a “provocational” tool as it gives rise to new questions and hypotheses, as well as insights about the text.

The current version of FeatureLens is developped at HCIL, University of Maryland, USA. We are using OpenLaszlo, Ruby and MySQL. An online version will be available soon.

Gamer Textually is a code artwork that generates visualizations and animations out of the text of the book “Gamer Theory” by McKenzie Wark. In the code, words become objects in a fractal space. Watch the animations and look at the large images to get a visceral sense of how the code ‘reads’ Wark’s work.

The code for Gamer Textually takes the form of a context free grammar which assembles artworks out of the original title, the table of contents, and the full source text of Wark’s “v1.1″ online manuscript. It also samples the cover logo art created by Tim Jones from the “v2.0″ print edition forthcoming from Harvard Press in April 2007. The code further imagines beyond Wark’s literal text, supposing myriad other combinatoric possibilities implied by Wark’s use of ‘leetspeak’, or letter-number substition.

Gamer Textually was created around one central question: How might a computer read Wark’s “Gamer Theory”? The answer takes the form of procedures or unit operations such as alphabetizing, ordering, coloring words by frequency of occurrence, etc. etc. However unlike a tag cloud or a histogram, which represents the statistical facts of a text, each Gamer Textually visualization represents one play session through the text – a series of chance encounters with various possible words structures, sensitive to initial conditions, always different and yet always the same.

Gamer Textually was written using Context Free, a cross-platform, open-source tool for visualizing context free grammars. It also builds on and synthesizes many contributions of the Context Free online community, including fonts and fractals featured in the credits. Work on Gamer Textually also inspired the development of a more general framework for digital text art, ConTextTree, forthcoming as an open source tool.

Gamer Texually in action

1. Contents and Alphebetization

One significant pattern that is machine-readable is the alphabetized table of contents. Why not a cover/title/table of contents that emphasizes both the structure of the work and the fact that it is ordered, like a file system directory?

2. Titles and Random Substitution

Gamer Textually:
1337title (AAA)
[image]
[movie]

I found it interesting that the manuscript I was familiar with, “GAM3R 7H30RY,” (version 1.1) was appearing in book form as “Gamer Theory” (version 2.0). There are many reasons why this might be the case. Perhaps a not-insignificant one is the fact that the stereotypical hacker substitution schemes of 13375p34k (“leetspeak”) intentionally obscure the contents from normal readers, and further make it difficult for a book to be Googled or found on store shelves. Let’s see, what was that book called… Gam3r The0ry? 6amer 7heory? 64m3r 7#30ry?? Of course, for a computer, complex substitution schemes are a bit easier to read and write, so perhaps a computer visualization of the title could encompass ALL the potential titles, randomly choosing different letter combinations each time….

3. Letters and Algorithm

Gamer Textually:
Handtitle (AAB)
[image]
[movie]

Gamer Textually:
Handtitle 2 (AAA)
[image]
[movie]

What about the letters themselves? What if they occurred as computational processes, not given, but emerging each time as a unique instance?

4. Space as Structure

Gamer Textually:
Ringtitle (AGD)
[image]
[movie]

Why not have the title and contents act like a structured and mutable object – like an element in a game, with spatial relations that implied not just a design space, but a play space?

5. Text and Tag Clouds

Gamer Textually:
Sierpinski Tags (AAE)
[image]
[movie]

At this point I was ready to start engaging the text itself. I used word frequency analysis to create weighted lists of all the words in the text, as well as ordered lists of the top 100, 50, 25 etc. Rather than display a traditional tag cloud, with one instance of each word sized accordingly, I decided to fill the space using a fractal, coloring the words according to weight but otherwise allowing random selections and repetitions to fill the smaller and gaps.

6. Logos and Animation

Gamer Textually:
Logotitle (BLT)
[image]
[movie]

The small logo from the front of the book design started to gain a new resonance for me – if the title was a structure, and the tag cloud was a space, then that small avatar might expect to enter Gamer Theory and interact. But this is a visualization, not a video game – I added an algorithm that randomly placed a variation on the logo at any one of many randomly selected angles of attack, depicting a computationally imagined superset implied by the one data point provided by the cover.

7. The Fractal Text

Gamer Textually:
Crystaltags (AAT)
[image]
[movie]

Gamer Textually:
Curl tags (ABM)
[image]
[movie]

Integrating the idea of the title and logo as objects with the idea of the text as fractal space creates a unified image. Each random seed renders a completely different visualization, yet for all their differences they are all, in the end, the same.

About the Artist

Jeremy Douglass is an artist and scholar in the field of interactive fiction and electronic literature. His current projects explore the poetry of context free grammars, folders, passwords, and tag clouds. He is a Ph.D. candidate in English Literature at the University of California at Santa Barbara, and writes for the text art blog Writer Response Theory.

Credits and Licensing

Gamer Textually images and animations are available under Creative Commons Attribution Non-Commercial 2.5 licence – the same terms that restricted the artist’s original use of McKenzie Wark’s text under the Visualize This! remix event. To the extent that these images and animations are considered original remixes or fair use sampling, the artist disclaims all rights and releases them into to the public domain, useable for any purpose whatsoever.

Gamer Textually images and animations were all produced using Context Free. The final form draws on code libraries helpfully bundled with the application and generously shared on the Context Free gallery and forums, including:

• The included i_pix, snowflake, demo2
• Block Typeface by Bargonaut
• Lowercase Alphabe by dzeni
• AlFreeBet by Guigui
• Sierpinski in the city by mizukinakeshu
• tritree

“The most comprehensive image buggery on the web.”

Toogle is a project that takes images from Google Image Search and creates ASCII-like versions of them using user submitted text. I also then put together images of the games discussed in each chapter of the text, using the chapter text as the content to make up the image.I also made a couple of images of a classic Super Mario Bros level using the full text.

http://c6.org/
http://c6.org/toogle

Toogle images

The Sims
Civilization III
Katamari Damacy
Grand Theft Auto
REZ
State of Emergency
Deus Ex
SimEarth
Mario Bros. (using the entire text of GAMER THEORY 2.0)
Sonic the Hedgehog (using the entire text of GAMER THEORY 2.0)