This entry will feature the notes for my presentation at ADRI to take place on October 17.
For those attending the workshop, you can download the tutorials here:
The notes for the presentation will be available soon after October 17.
Josh On, They rule: http://www.theyrule.net/
The following are the notes of my presentation, part of a workshop at ADRI on October 3, 2017. They are made available on Remix Data for those who attended the workshop, and for anyone online who may be interested in data visualization as a creative tool. The notes are meant to function as an entry point for anyone who is not familiar with data visualization, but hopefully, even those who are well acquainted will find relevant some of the concepts, and especially the examples discussed. You can download the technical tutorials of SVGs and a basic bar chart developed with D3, which were discussed at the end of the workshop:
This session focuses on how data visualization relates to creative practice. First I will go over the basics of data visualization to then provide a few examples of artists, designers and researchers who have implemented some form of data visualization to produce creative works of art or design.
Three basic media forms to analyze:
No matter which type of form is analyzed, the data usually will be organized as textual data that can then be visualized and represented in some way.
The data in turn can also be represented in the three forms
Image (data visualization)
There are many tools available for analysis at this point, for this reason one should become familiar with as many related to one’s interests before deciding to engage in advanced development.
Owen Mundy: https://owenmundy.com/
One thing to keep in mind is that information based production, which is the platform that informs our way of working at the moment, consists of key stages, which include:
One thing that I find Digital Humanists should be aware of is of becoming too focused on technological innovation. In effect, one may feel certain pressure, in order to become an established researcher, to develop something that appears technologically innovative as opposed to using tools that may already be available. This could lead to a model of innovation for innovation’s sake, which could be paralleled to moments in the history of art, in terms of “art for art’s sake.”
Brooke Singer: http://www.toxicsites.us/
The key thing, then, is to begin with questions that are of real interest to a researcher, and based on those questions, one should search for tools that will be useful. Eventually, one may develop a need for specialized tools which are not available, and that is when one may become an actual developer.
To drive the point home on this issue, in the spirit of remix, I borrow a quote from Charlie Gere, who in turn borrowed it from Gilles Deleuze: “the machine is always social before it is technical. There is always a social machine which selects or assigns the technical elements used.” 
If one has never had exposure to Digital Humanities methods, then one may not find how it can be relevant to one’s research. So, one should understand that the questions that may already be informing one’s research can be reshaped once one is exposed to different tools. With this in mind, I will explore some questions that I usually introduce to students who may not be familiar with text-mining or image-mining.
Key Questions for all analysis (image, sound, text):
The above questions have embedded the foundational question for any subject: How does “it” work?
To engage with this basic question, one usually will analyze the material (image, sound or text) carefully, then writing down one’s observations, which in theory, others could review and analyze. With data-mining tools, one can actually use quantification to evaluate patterns, to then consider how such patterns may shed light on one’s question, which often times leads to more complex questions on the subject.
This process then consists of four parts:
• Perform research based on a question (image, sound, text)
• Gather data
• Evaluate data
• Represent the data (image, sound, text)
Giorgia Lupi and Stefanie Posavec, Data drawing pen pals, http://www.dear-data.com/
Evidently, this process cannot be covered in one session, so what I will show at this point are a few data visualization examples for image, sound and text, so that we get a sense how this process takes place. Given that this session is about data visualization, I will finish by focusing on the development of basic shapes that can be used to visualize any type of data. This might be basic for people acquainted with web development and programming, but the point is to keep in mind the conceptual process that informs how and why we come to choose certain forms to represent/visualize our research.
Artists and designers have actually used data visualization to produce creative work. Some examples include (discussed with participants during live session):
Sound represented visually:
Manny Tan & Kyle McDonald,
Dear Data MoMA Collection:
Entry explaining the process:
Data drawing pen pals
In the spirit of the above examples, particularly the last one, I will focus on basics, because, as you can see, we don’t have to program to begin visualizing information. We can draw our interpretation of data. It is, however, important to eventually program your visualization so that it can reach a fuller potential. What is worth noting here is that when we visualize data, we are actually interpreting it. Data visualization can be a fair presentation of raw data when we consider how and why we choose colors, line shapes, spacing between objects etc. Such basic graphic objects affect how the data is perceived. We must be conscious of this aspect of data visualization, and make sure we are being fair to the project when we decide to use specific visual elements to present factual information.
Data Visualization is in large part developed with SVGs. For this reason, I will go over how basic SVG shapes are developed. It is my experience that having a basic hands on understanding of the foundation of tools one will use helps one develop an intimacy that otherwise will never emerge if we only function as users. In other words, knowing what the code behind tools actually does helps in deciding how to use such tools.
You can download the tutorial if you don’t know how to use SVGs by clicking the links below. A basic bar chart is also included which can help in developing an initial sense of how data viz functions.
 Gina Neff and David Stark, http://www.umass.edu/digitalcenter/events/2002Workshop/Papers/stark_permanently.beta.pdf, accessed October 2, 2017
 Charlie Gere, Digital Culture, “Introduction,” Digital Culture (London: Reaktion Books, 2008), 17.
After the film’s screening, during class, we usually discuss the types of shots used, and their timing in relation to sound. One thing that becomes evident when visualizing the film as a grid-montage (figure 1) is that there are a lot of close ups. It is therefore worthwhile to evaluate the pervasiveness of the close-up in Marker’s film in order to understand how it is used to tell the story of a time traveler.
As we can see from the grid above (figure 2), most of the film is made of close-ups. These shots could be divided into various categories but for this brief analysis, I chose shots that included a single person’s face in a frame. There are other close-ups of hands and text that repeat, which are also included for a general count, along with occasional shots of two persons if they offered a more or less abstract environment. There are about 135 close-ups in the film. And they become predominant once the basic environment in which the story takes place is established. Based on this Marker organizes the film into three sections. The first part set-ups what happened in Paris, and how humans had to go underground to survive. Establishing shots of this major event is combined with shots of an airport, where a mysterious murder is witnessed by a woman.
The basic structure of the film for the first two sections is that of establishing shots and mid-shots moving into close-ups. After this the film goes back and forth between time realities using action shots or close-ups to create psychological tension. There are a few establishing or whide shots during key moments, such as when the hero and the woman visit the museum. But close-ups, action shots and wide shots will not be combined until the end of the film, when we find ourselves in the airport, which, as we learn at this point, is also the beginning of the film. And we come full circle.
Based on this general analysis which could expounded in much greater detail for each section of the film, we find that the close up is effective for Marker because of two reasons.
1) It displaces the environment in which the story takes place and this pushes the viewer to engage with the characters directly as people.
2) On a pragmatic and practical level, it is economical in terms of a budget because little expense goes into the set.
The results of this is tension and drama that is built through affect, meaning that we are pushed to identify with an image in abstract fashion based on empathy and/or sympathy.
The last third of the film may be hard to follow because there are no wide or mid shots during the hero’s final debriefing. The viewer bluntly confronts a series of faces that stare at each other, or sometimes straight ahead–just off camera. There is much more to be said about La Jetée‘s structure, and why it is considered one of the greatest films ever, but one of the reasons why this film is historically important is due to the fact that it makes the most of the close-up, or affect image, to develop a sci-fi psychological thriller that defies film genres to this day.
Figure 1: Visualization of 30 Music Video mashups taken from YouTube. Mashups were uploaded between 2006 and 2011
I am currently finalizing research on 30 YouTube music video mashups, uploaded between 2006 and 2011. I started this research during the last days of my post-doctoral work at the University of Bergen in May of 2012. I was able to refocus on this project during 2015 thanks to the support of a Faculty Research Grant from The College of Arts and Architecture at The Pennsylvania State University. I thank them for their ongoing support. In what follows I will introduce the principles of this research project.
Figure 2: Preliminary Visualization of the first 10 of 30 music video mashups. The list includes from top-left to bottom-right: 1) Danger Mouse Encore (Grey Album, Jay-Z and Beatles), 2) Radiohead vs. Dave Brubeck, 3) 99 Problems with Buddy Holly, 4) Ray of Gob (Madonna vs. Sex Pistols) , 5) The Strokes vs. Christina Aguilera, 6) Eminem Featuring Aerosmith Sing On For The Dream Moment, 7) Blondie Vs. The Doors – Rapture Riders, 8) Boulevard Of Broken Songs – Green Day ; Oasis ft.Travis ; Eminem, 9) Sour Glass (Portishead vs. Blondie vs. Kanye West), 10) Weezer vs. Queen.
Color code: light-green = special effects or titles, blue = main-track/instrumental, yellow= lyrics/complementary track, red = footage of both videos combined.
During my postdoctoral research I worked on three case studies of YouTube video memes. I published my findings in the paper “Modular complexity and remix: the collapse of time and space into search” Based on the editing patterns that I noticed the YouTube video memes shared, it became evident to me that a particular grammar and syntax was at play in the way people edited diverse forms of music video remixes; this grammar also appears to be linked to the way music mashups are produced and function, and for this reason, after I finished my research on the music-video memes, I decided to look further into video-music mashups. I chose 30 of the most popular mashups between 2006 and 2011. Given that remix as a form of cultural production is closely linked to basic music remixes it is worth evaluating how meme patterns move across different genres that combine image and sound as a single form of expression. Music video mashups are good case studies for this reason.
To analyze the patterns of the memes more accurately, I realized that it was not enough to visualize a grid of the video footage as I had done for my previous research. I needed to see how the footage of different videos was edited as one to complement the actual music remix. Therefore, I developed a color-coding system to mark the video edits. I did the first 10 manually, by laying color over the video montage (Figure 2).
This process took too long and was not very accurate, so I developed a batching method in which I evaluate the frames that correspond with the original music videos and assign them a numbered sequence that can then be run to create a montage that matches the original video montage. I share the examples of the first three videos below.
Figure 3: “Danger Mouse Encore (Grey Album, Jay-Z and Beatles)” Grid-montage of music video mashup. Video available at http://www.youtube.com/watch?v=qFd1QvNPKmI
Figure 4: “Danger Mouse Encore (Grey Album, Jay-Z and Beatles)” Video color-coded pattern. Color code: yellow = special effects or titles, blue = instrumental/main-track, green = lyrics/complementary track, red = footage of both videos combined–or hybrid video with special effects.
Figure 5: “Radiohead vs. Dave Brubeck” Grid-montage of music video mashup. This version is no longer available on YouTube, but an alternate remix can be found here: https://www.youtube.com/watch?v=0PQQPUb8Rl0
Figure 6: “Radiohead vs. Dave Brubeck” Video color-coded pattern. Color code: yellow = special effects or titles, blue = instrumental/main-track, green = lyrics/complementary track, red = footage of both videos combined–or hybrid video with special effects.
Figure 7: “99 Problems with Buddy Holly” Grid-montage of music video mashup. Video available at: https://www.youtube.com/watch?v=4bUAzC9wgkw
Figure 8: “99 Problems with Buddy Holly” Video color-coded pattern. Color code: yellow = special effects or titles, blue = instrumental/main-track, green = lyrics/complementary track, red = footage of both videos combined–or hybrid video with special effects.
Based on these three videos we can see that the approaches vary but nevertheless there are some similarities already evident. For one thing, there is a need to go back and forth between footage in order to reinforce the fact that the video is of a music mashup of two songs. The first two videos show some aspect of the video (red) that supports the aural experience of the sound mix colliding. This is also evident in my preliminary visualization of the first ten videos (figure 2). There is an exception of two videos (mashups 3 and 5), which don’t have any combined or hybrid footage (red), but only juxtapose material of the originating sources. When listening to the actual music remix, it becomes evident that it is because of the way the two songs are mashed that the videos are edited with basic montage. In all of the videos I have analyzed thus far, the image appears to support as best it can the aural experience of the remix. The reasons behind this begs a detailed theoretical examination. It is my aim that my analysis will provide a good sense of how image and sound complement each other in music video mashups. I have looked at more recent mashups on YouTube since 2012, and find that the elements in the 30 videos I am analyzing are still at play. This may mean that certain aesthetic and expressive conventions of communication are already well in place among video and music remixers. In the future, I will be sharing more details of my findings on all 30 videos in a formal paper.
Figure 1: Detail of Minima Moralia Redux Remixes 51 – 55. First set of entries part of the second part of Minima Moralia Redux.
Note: This entry was updated on April 19, 2015 in order to add details on formal aspects of the project.
Minima Moralia Redux, a selective remix of Theodor Adorno’s Minima Moralia, enters a second phase in 2015. This was not foreseen when I began the project back in 2011, because the work is not only a work of art, but also research on data analytics, as well as a critical reflection on networked culture.
The first part of Minima Moralia Redux (entries one to fifty), consisted of updating Theodor Adorno’s aphorisms–that is to remix them as contemporary reflections of the way global society and culture is engaging with emerging technology. When I finished the first section, I realized that the project’s aesthetics were changing. This was for a few reasons. In terms of research, the first section provided more than enough data for me to data-mine Adorno’s approach to writing; therefore, I came to see no need in following this methodology. I plan to make my findings about this aspect public in a formal paper in the future.
Figure 2: “of the truth comes,” part of a sentence of Minima Moralia’s aphorism 54 that, when clicked, opens a Google search with relevant results.
In terms of art as a form of reflection on the times it is produced, it became evident to me that Adorno’s writing needed to be connected directly with the network on which it functions as a remix; and for this reason, I opted for the current format, which consists of the text as it is available in English at Marxists.org, translated by Dennis Redmond.
Figure 3: “Truth,” part of a sentence of Minima Moralia’s aphorism 54 that, when clicked, opens a Google search with relevant results.
The format for part two of my art and research project consists of linking phrases, parts of phrases, or just single words, devoid of any punctuation, to corresponding Google searches. (Figures 2 and 3) The reader can click on any link and be taken to respective search results, which will change according to Google’s updates on its search engine. The concept of making every word in a text a link is certainly not new. Heath Bunting had already explored this concept with his early net.art piece called readme (Figure 4), in which he took every single word in a review of his own work published in The Telegraph, and linked it to corresponding .com URLs, some which did not exist at the time, but Bunting’s proposition was that one day they likely would.
Figure 4: Heath Bunting, “Readme,” (1998). Online artwork that links all words of a Telegraph review of Bunting’s work to specific .com URLs.
The second phase of Minima Moralia Redux functions more or less in the same way as Bunting’s work, only in this case, the words or phrases are linked to search results as opposed to specific URLs. In this way, Adorno’s second section of his book becomes interconnected with an apparatus (networked communication) that he likely would have been quite critical of. The second part of Minima Moralia Redux, consequently, is an intertextual mashup of Bunting’s early net.art piece and Adorno’s writing.
The above is made possible based on complementary technical and formal implementation of data-visualization. To develop visualizations in support of the new format of links to Google searches, I decided to keep using word-cloud visualizations, at this point Wordle (I used Many Eyes in the past), which provide an emphasis on certain terms for each entry. The number of words vary, and may range from three to as many as sixteen. These words are then used to perform an image search on Google. I then take a screenshot of the top results and place it at the top of the entries. I also input the text for each entry on Voyant in order to see the overall word frequency and their contextual relation (I don’t include my qualitiative analysis of the entries at this point, but it is important for future in-depth evaluations of the project). The text-mining results, along with the word-cloud visualization are placed at the bottom of the page to share the sources used in the development of each blog entry.
Beyond these formal and technical aspects of the project, there is another layer in the second phase of Minima Moralia Redux. It also borrows from Traceblog a previous blog project which I finished in 2013. In Traceblog I made public, ghost logs of my searches online. The logs were fake results produced with a free plug-in for Firefox called Track me not. In Traceblog I wrote nothing, but rather made available material that was produced based on the “disguising” of my online activity. The second part of Minima Moralia Redux is similar in that I don’t produce anything, but rather repurpose pre-existing content, in this case Adorno’s aphorisms, which, when put through a search process, give way to showing how search is connected to content online. This type of writing is more a form of curating the sentences, phrases, and words to provide searches that do not lead directly to Adorno’s book available on Marxists.org, but to the phrases that form Adorno’s argument as they may appear on thousands of websites at divergent times.
This led me to realize an important aspect of sampling in general that I plan to develop further in theoretical essays, but anyone who has followed my argument thus far will find evident in this brief reflection, which is why I think I must elaborate on it at this point.
What the linking of phrases or words for each entry in the second part of Minima Moralia Redux makes evident is how we come to develop what we may consider original content. I realized that if I put an entire sentence into a google search I would have Adorno’s publication at the top of the search, which meant that Google was able to recognize the string of terms as a direct “sample” from Adorno’s writing. I wanted to get diverse results that did not lead to Adorno directly, so I decided to adjust the search to a string of words that would provide a result that would not bring Adorno’s text within the first page of results. This at times is not possible, but what becomes evident in this process is that we develop our own work from things that are pre-existing. A single word search is likely to provide the most diverse result on Google, but the more specific the string of words, the more likely one is to reach a specific “sample” that may be deemed the original work of a particular person.
In effect, this second phase of Minima Moralia Redux exposes that what we tend to recognize as someone’s creation in any media is really a specific combination of elements that are mashed together by producers according to what they want to communicate, or express. In effect, even when we speak, we are borrowing from a set of samples (words) archived in a database we carry, called our “memory.” Such combinations are seen as “property” when they are placed in a format that is more static–a product. With the speed of network communication, however, the static state of things is coming to an end, and the ever-changing state of forms produced (viral memes are an early example of this) will become valued more than a single instance of production.
Nothing is original, just unique.
Figure 1: Overall view of a force layout visualization of Theodor Adorno’s Aphorism one, Minima Moralia, developed using D3, Eduardo Navas, February 2015.
Note: For the first post on Adorno’s text, see “Preliminary Notes on Analysis of Theodor Adorno’s Minima Moralia” part 1.
During the month of September of 2014, I began evaluating different types of visualization methods which could be used to examine the quantifiable relationship of image and text in closer relation to semantics. This can be a challenge for researchers who are using quantification to analyze visual and literary works, which in the past were solely evaluated with comparative methods relying on close readings of the subjects of interest.
My decision to develop D3 force layout visualizations of selected aphorisms of Theodor Adorno’s Minima Moralia came about in part due to extensive discussions on research methods I have been having since the beginning of the 2013 academic year with Graeme Sullivan, Director of the School of Visual Arts (SoVA) at Penn State. At one point Professor Sullivan wondered if a tool could be developed or repurposed that would show the relationship of different elements in clusters. He was interested in using such tool for studies in Art Education and related fields of research. I thought that, in my case, I could use it for visualizing different projects of my own.
In effect, I evaluated D3 and realized that its force layout features could be effective in developing visualizations for the multiple purposes I had been considering for over a year. In my case, this meant an in-depth understanding of Theodor Adorno’s writing for my project Minima Moralia Redux.
In what follows I explain how I will be implementing the D3 visualization template that I developed in order to write a critical analysis of Theodor Adorno’s Minima Moralia as part of my project Minima Moralia Redux. The visualization I share in this entry is the first of six that I will be developing in the next few months.
Figure 2: Visualization detail highlighting the number of recurrences of the article “the” throughout the text.
Prior to the force layouts, I had been using Many Eyes to visualize the importance of words in the aphorisms. [figure 5] These visualizations were not meant to provide a quantifiable presentation of the number of words, but rather offer at a glance an idea of how the aphorisms tended to have certain words repeat more than others, and consider this in relation to one’s reading of the actual text. It must also be kept in mind that Minima Moralia Redux is an actual artwork, so the quantification of information is used in part for aesthetic exploration. At the same time, I aim to develop a precise understanding of Adorno’s writing approach. One could argue that there are some limitations to my project given that I am only analyzing the English translations of Adorno’s book, not the German version. Nevertheless, doing the analysis of the English version available online in comparison to the official Verso English publication, in my view, is a valid pursuit, given that most people will be exposed to these versions of Adorno’s writing, due to the prevalence of English around the world. For this visualization, I used the online version of aphorism one.
The force layout is able to expose certain elements that a word cloud is not able to make evident. For one, it shows the recurrences of each word with its actual number. This was deliberate on my part. I could have shown only the word with a circle corresponding in size to the word’s recurrence. This means that , if I had not included the actual number of recurrences, in a way, the force layout would function similarly to the word cloud–by giving a somewhat abstract idea of the relation of words. But I decided to include the actual number of recurrences because I want the size of each circle to be equated with a specific number that can then be compared with other circles and numbers in the visualization. The connections among the words, in effect, are developed in relation to the recurrence of each word, which is interlinked to the words that come before and after. The result is a visualization that provides a sense of the actual relation of vocabulary to repetition, and how this might play a role in the argument of the text itself. what follows explains how this takes place.
Figure 3: Visualization detail highlighting the number of recurrences of the preposition “of” throughout the text.
Word clouds show that the most common words will always be articles, prepositions, and common verbs. These are the binders in a sentence. Without them, we would not be able to communicate as we currently do. Many word clouds, such as the one available on Many Eyes and Wordle, omit common words. The results are visualizations that give a vague but decent sense of what the content may be about. What is not evident is how the vocabulary at play is enacted to make the argument effective.
With my implementation of a D3 force layout visualization, I aim to examine a few things about Adorno’s approach to critical writing. For one, the particular entry being examined here has three words that are mostly used, which are usually among the most common in all texts: “The” (24), “of” (19), and “is” (14). (See figures 2-4.) The number of words repeated after these diminishes drastically. Here is the list of the top ten words that recur more than once, which I extracted using Voyant:
For a complete list, access the text on Voyant.
The total number is 368 words and 212 unique words. It must be noted that not all the words in the text were visualized, only the words that repeated at least twice. This was done for the sake of simplicity. If all words had been included, the visualization would have been completely interconnected except for two words, because all the words would be linked to the terms that come before and after, thus creating a hard to read visualization.
What we have instead is a force layout that shows how the most repeated words bring together other words to develop a text that is dense in content but minimal in actual repetition. The first hint of this is the number of unique words, which is 212 out of 368–that’s roughly two thirds of the text is composed of unrepeated terms.
Figure 4: Visualization detail highlighting the number of recurrences of the verb “is” throughout the text.
Based on my assessment of this visualization in relation to the repeated and unique words, it appears that a text that makes use of a large vocabulary will show a few words that are used only when necessary to connect unique words. Meaning, an essay that is redundant will result with a greater number of large circles, differing from the visualization of aphorism one, as shown in this entry. The implication of this is that Adorno’s aphorism was carefully constructed, given that it would take great effort not to repeat words while trying to develop a complex argument. Trying not to repeat words for the sake of not repeating them would be an empty exercise, so it must be kept in mind that the high number of unique words comes at a rigorous process of selecting from a large vocabulary in order to produce the best argument possible.
This, however, does not automatically mean that a text with less unique words is not as effective, or unable to convey meaning. What can be assessed from this brief analysis is that Adorno took the time to think about the composition of his text, and that he did actually practice what he promoted in his critical writings. This could be perceived when a person reads the text, of course, but the numbers and visualization make this evident with concrete data. The entire premise of Minima Moralia, in fact, is to be critical of mindless recursion. Adorno believed that one should be wary of repetition as a regressive mode leading to passive thinking. As the force layout visualization shows, Adorno’s approach to life was implemented into the formal development of his work.
Figure 5: Word cloud of “For Marcel Proust” as visualized for Minima Moralia Redux, Aphorism 1 Remix.
There are more details that could be discussed about the relation of words to the actual content of the text, such as the numbers of the words to the actual message being conveyed. In this case, Adorno is reflecting on the importance of Proust, and Proust’s name is only mentioned once, as part of the title of the aphorism but never within the actual text. One would never be able to assess Adorno’s views on Proust by analyzing the visualization alone. This means that one must always be engaged with the actual text, and consider any type of visualization supplementary to the experience of the content. The quantification of data can be helpful only in verifying what one may perceive is at play within the aesthetic experience of a work of art. This, I find, appears to be a constant challenge to individuals who quantify data to evaluate aesthetics, which may appear like a paradox.
Figure 6: Visualization detail highlighting the number of recurrences of the word “one” throughout the text.
Others things that I will be evaluating once I visualize the remaining six aphorisms I have selected is how word usage may fluctuate from entry to entry over the development of the actual book and how this may relate to the number of unique and repeated words in relation to the overall focus of the work.
The potential of a D3 force visualization in terms of general use is to understand how the size of words or images (this type of visualization can be used for image data as well) can become an effective mapping tool, once a standardized relation of the nodes created are understood with a specific meaning consistently. In this case, this means that a few large number of circles versus a large number of small circles implies diversity in vocabulary. For this to happen, however, parts of the process need to be automated, so that the input of information is much faster.
Figure 7: Visualization detail highlighting the number of recurrences of the word “play” throughout the text.
Figure 1: Eduardo Navas, #Subaltern, 2010 tweet rewritten as a poem, February 2015
Selected Poems in D3 Force Layout
#Nature || #Opinions || #Fatty || #Subaltern || #Migrants
#Modules || #Nano_specs || #Standards
#Abundance || #Plastic || #Universals || #Predominance
During the month of January and February of 2015, I began to consider how to reconfigure selected tweets of my @poemita twitter account as poems. The first outcome of this process was three sets of image-layouts of selected poems from the years 2010-2013 which I called “Poem Portraits.” They are available on the main @Poemita project page:
Simultaneously, I had been working with D3 to develop a force layout for visualizations of selected entries from my project Minima Moralia Redux (This set of visualizations will be discussed in a separate entry). Such layout is designed mainly to show the relevance of words within each of Adorno’s aphoristic essays. At one point in this process, it occurred to me that I could use D3 force layouts not only for research based visualizations, but as an actual medium to rewrite poems. Hence, I repurposed D3 features to develop a set of poems as shown in figures 1 – 4.
Figure 2: Eduardo Navas, #Standards, 2012 tweet rewritten as a poem, February 2015
Some of the poems I selected also are part of the Poem Portraits, but the interactivity that D3 offers led to a very different feel for each piece of writing. The basic premise behind the force layout poems is that each line of text should be connected to the line below it and above it, so that each poem could be read similarly to a print version. One can get a sense of this when comparing the D3 layouts to the Poem Portraits.
Figure 3: Eduardo Navas, #Abundance, 2013 tweet rewritten as a poem, February 2015
D3’s force feature allows for the words to push away from each other, and it is the lines that keep the words linked; otherwise they would float away randomly. These simple features offer unexpected results, which make the reading of the poems potentially different every time they are accessed. There is also a sense of discovery as the reader needs to figure out how the words are connected. Some poems are move ambiguous than others, and one could read the lines in different order.
Figure 4: Eduardo Navas, #Predominance, two 2013 tweets rewritten as a poem, February 2015
I found that poems with longer lines took on a layout similar to their print counterparts, while poems that had more unconventional formats turned out more ambiguous. #Subaltern (figure 1) and #Abundance (figure 3) for instance, read very much like their corresponding print versions. Nevertheless, when each poem is launched, the first line could begin at the top or bottom of the page; also the layout could be reconfigured by dragging the circle with the mouse to any area. This offers some play with the layout of the words. And as it is well known, the layout of words is always of out most importance to poets.
#Standards (figure 2 ) and #Predominace (figure 4) offer more ambiguous layouts. Each poem may take longer to read given that the words are connected in ways that make sense to the corresponding print version, but due to D3’s random feature at initial launch in combination with its force feature, the reading of the lines turns out to be open-ended.
At this moment I have a set of tweets from the year 2014 that I am working on, which likely will be influenced by my awareness of the creative potential behind a force layout and a more conventional print layout. I’m not sure how the next set will turn out, but the experimentation with the years 2010 -13 is definitely informing the ongoing rewrites of tweets.
The image below is a grid-montage visualization of Godard’s Breathless. It was developed from a dutch-subtitled version of the film. I will be posting a detailed analysis in the future. In the mean time, this preliminary visualization is shared for research purposes as it may be of of use for anyone who wants to attain a general sense of the scene sequence, camera angles and frequency of shots throughout the film. I developed this and other visualizations (to be shared in the future) in order to teach film/video history, theory and editing in terms of art practice.
It should be noted that the frames of this visualization are not accurate to the actual shots. They were generated to attain a general sense of the scene sequence. Some important shots may be missing due to the fact that I exported in-out shots sensitive to major movement in the entire film, which then were sequenced in increments of three (which means that some in and out images are missing). The visualization is meant to provide an overall sense of the film’s development, not a detailed examination of the montage. This is crucial to keep in mind with a film such as Breathless, in particular, because of the importance of the jump cut.
More will be explained in a future and detailed post, which will include other visualizations and data analysis.
Figure 1: selected shots from Capote (left) and In Cold Blood (right).
Interdisciplinary Digital Media Studio is a class in the IDS program in The School of Visual Arts (SoVA) at Penn State in which students are introduced to methodologies and conceptual approaches of media design. For the class, I taught them how to research and develop design presentations with the implementation of data analytics for moving images and texts.
One of the assignments consisted in analyzing the films Capote (2005) directed by Bennett Miller and In Cold Blood (1967) directed by Richard Brooks in relation to their corresponding books, Capote by Gerald Clarke and In Cold Blood by Truman Capote. We viewed the films in class, and read, both, the novel and the biography. The class then analyzed the respective books by doing word searches, analysis of specific passages, and creative approaches by the respective authors, to then evaluate those searches in relation to the films. For the films I provided montage visualizations, which are selected screen shots representative of all the scenes (figures 2 and 3).
Figure 2: Visualization of the film In Cold Blood (1967). Click for a larger image.
The students were free to use the film visualizations and the data-mining of the texts according to their interests. Some of the options I suggested, but which were certainly not the only possibilities, included: times names or any other terms were mentioned, difference in the way particular moments appeared in the films and the books, and which parts were omitted.
Figure 3: Visualization of the film Capote (2005). Click for a larger image.
The students had to design an infographic, or a similar form of visual design that presented their findings in an easy to understand format. The students had to also write an assessment of their findings. What follows are selected projects which I consider successful for elements that I explain for each case. All my students worked very hard on their visualizations, and ideally, I would like to show all of their projects, but for the purpose of this post, I selected designs that show unexpected or experimental approaches in the implementation of analytics to find possible answers to diverse questions on the same subject. The samples that follow are successful in some ways, while having shortcomings in others.
Figure 4: Visualization of common elements between Capote, the book and its respective film adaptation, by Michael Regan. Click image for larger file.
Michael Regan looked at commonalities and differences in Capote the book and film. In this regard he writes:
The focus and content of my infographic is in the arrangement of common elements between a book and its film adaptation. This is a very relevant way of viewing the comparison of two media representations of the same content. It is very interesting to view how differently the film creators had to move content to best adapt the form. The In Cold Blood film rearranges everything from the book into a different order to better fit a film format, while the Capote film takes most of its content entirely from one section of the book. These are two ways of adapting a book into a film, and looking at the infographics allows a very quickly visual way of understanding these techniques. It also shows the way that the books are constructed in the first place. The In Cold Blood book shifts back and forth often among the murderers, the victims and townspeople of Holcomb, and the investigation. This helps to explain the switching back and forth of elements and their ordering between the film and the book. This is contrasted by the writing style of the Capote biography. The small relative amount of time the book spends on the In Cold Blood writing process, as shown by the infographic, shows the focus of Gerald Clarke, the author. He chooses to explore the entirety of Capote’s life, without giving an undue amount of attention or priority to the In Cold Blood part of his life, however sensational a time it may have been.
Regan’s visualization represents the emphasis of the film on particular aspects of the book, and where they may or may not overlap. One can get a decent sense of the adaptation process, and to some degree, assess how such process is enacted in the particular production. His visualization is indexical, meaning that it allows the user to go to specific areas that Regan evaluates are important to the process of adaption from book to film.
Figure 5: Visualization of color mentioned in the two films and books analyzed in class by Drew Heo. Click image for larger file.
Drew Heo decided to focus on the role color played in the films and the books. In his assessment paper he writes:
Color is a visual component, and Truman’s novel and biography are black and white on paper. Due to the two differing mediums, it’s only fair to allow both works to shine in the fields they are best in. The bar graph is a visual indicator of the specific mentions of colors in the texts of both books, as well as the amount of times a color is verbally spoken in the films. As expected, the films fall short when compared to the writings, as color is a visual thing and not used as much in speech. In order to compensate for the lack of mentions, below the graph is a condensed form of the colors present in both films in a “Movie Barcode” format, which has become recently popular on the internet for analyzing films.
To consider how color plays a role in the film is a unique question certainly worth pursuing. The accuracy of the numbers do need to be double checked, but what is worth noting here is his particular approach to evaluate a concept across media that is not easy to quantify. If anything, his focus and approach demonstrate a potential for more in-depth analysis of details shared across works that are adapted and crossover relevant productions that are often compared, but hardly ever analyzed quantitatively.
Figure 6: Recursion of characters in the biography, novel, and respective film adaptions, by Ethan Jones. Click image for larger file.
Ethan Jones focused on the number of times the names of particular characters appeared in the text, and the number of times each character appeared in the film. He writes:
There was a decision early on to reduce the commonalities between all four works by omitting what was exclusive to one story or the other. For example, the Truman Capote character is omitted because of his lack of presence in the In Cold Blood story (book and film). On the other side of that, minor characters from In Cold Blood (book and film) were discarded because of their lack of involvement in the Capote biography. Several characters remain constant, such as Perry, Dick, Alvin Dewey, Susan Kidwell, and the Clutter family. Simply looking upon the amount of times a given character was mentioned proved to be sufficient.
With Jones’s barchart we can evaluate how the characters are at play in all four works. One thing to note is that he counts the appearances of a character in the respective films every 5 frames. This, however, is dependent on the selective shots I provided, so the number of times a character may actually appear is not related to every five frames from the actual film but frames based on my selection of shots representing each scene. Nevertheless, the chart does provide a general idea of the prevalence of the characters in the respective publications and films. This can be cross-examined with the analysis of other students which is not included in this entry.
Figure 7: Analysis of parts of the book and film where dialogue did not play a major role, by Nikki Tatsumi. Click image for larger file.
Tatsumi focuses on parts that had next to no dialogue in both the film and the book, In Cold Blood, she writes:
For this project I focused on the parts of the movies and books where dialogue did not play a major part in the creation of the scene. This was most relevant in the book In Cold Blood. This book, borderlining a research novel, was beyond descriptive. Its specificity allowed the reader to truly identify and imagine these people to the point of skin color, body language, tattoos and countless other tidbits that Capote decided to include. If not enough, Capote’s rendition of each character’s backstory also helps to fuel the reader’s imagination. It is through these scenes that I think Capote is the most successful. Though dialogue gives a character a voice, Capote was able to utilize his research and observations to create characters who did not need one. By their mannerisms alone, Capote could write this whole story. But when these heavily described sections of the book meld with the movie. Directors used the score to help make up for lost information through the text. In Cold Blood and its movie, had
the best transition.
In the movie version, heavy use of jazz influenced music helped to create the mood. It was a lot more foreboding due to the the choice of score. It was the type of music that was not soothing or collected, it was erratic and soft at times, loud and uneasy at others, as if the viewer did not expect what was next. The scenes that this score was utilized really well were the juxtaposed part of Perry’s past and the beginning scene with Perry at the bus terminal. They utilize the uneasy jazz to mirror Perry’s
feelings at both times.
In Tatsumi’s design we can note that she was selective of which parts to feature, as there were many more moments and passages in both the book and the film that would fit the subject of her research. In this case, we can consider her focus and design similarly to Heo’s, with potential for more in-depth exploration. Her focus also shows how analytics, while at times may be presented to be objective in what the results may be, are actually informed with particular questions and interests that drive the research. This is something that individuals using quantitative tools need to be more honest about, and not claim that what they found is entirely “objective.” Tatsumi’s focus is unique in that she focuses on moments in which dialogue did not play a major role. No other student considered this, and I never would have thought of this focus. In this sense her research points to an aspect of the film that could be very interesting to further analyze.
One thing that I should note is that her design could be much better. However, given that many of the students were dealing with multiple elements of creativity and research I did not push her to redesign her presentation given that she had touched on an interesting aspect of both the film and the book that she could have explore more in-depth.
Figure 8: Diagram of biography, novel and both films by Megan Coren. She focused on similarities of the four works. Click image for larger file.
Megan Koren focused on the similarities of all four works. Similarly to Tatsumi, her diagram shows that she was rather selective. She writes:
The nature of my infographic does not show many of the differences between the books and films; rather, it shows the similarities of In Cold Blood and a follow-up comparison of how Capote visually paralleled this information. Because Capote was a secondary addition, I feel I provided a skewed representation of the biography. I included text from the book that aids the comparisons in my infographic, but Capote is not really about the novel In Cold Blood. It is about Truman Capote’s life, spanning childhood through adulthood. Although interviews with the Clutters cover only roughly 50 pages of the monstrous 547-page biography, the film diminishes this biographical time-line and completely focuses on Capote’s writing of the book. The film concentrates on the obsessive Truman himself, while shot-wise it accurately parallels the In Cold Blood film. We see the similarities between Capote and Perry and why Capote sympathized with him so strongly. Regardless of their different perspectives, I was genuinely impressed at how many connections the directors were able to make between scenes in the Capote and In Cold Blood movies.
This particular presentation is easy to read, and one is able to figure out how the four works are related according to Koren’s interest. However, it does have inaccuracies that are evident upon closer observation. for instance, the moment when Perry speaks on the public phone is not at 0 minutes of the film, but actually more or less around 10 minutes into the film. Koren should have spent more time with the films to evaluate their respective chronologies. If she did this her assessment, which is quoted above would be stronger. Her selection of passages from the books also lack a systematic approach, which makes one wonder why the citations were chosen over any others. In any case, Koren’s design shows potential in exploring formal aspects of presentation of a large amount of material. She needed to spend more time working on the accuracy of the data and information–her design, conceptually, shows potential, although it could be further polished for a more elegant read.
Figure 9: Visual analysis by Caleb Yoder of the names “Dick” and “Perry” in book publications Capote and In Cold Blood, and their respective film adaptations. Click image for larger file.
Caleb Yoder decided to be very specific in his analysis and focused just on the names of the two killers. He writes:
After data-mining information for Dick and Perry, I found that the books are often able to provide a lot more in terms of description for the characters. On page 325 of Capote, Clarke is able to give a thorough description of Perry’s appearance, personality, and childhood. The films need to rely on the actions of the characters or their conversations, and in some cases even flashbacks (like Perry’s multiple flashbacks to his childhood in the film version of In Cold Blood), in order to convey this information that the book lists off so easily. Film must also take some creative liberties with how the story functions visually–the book doesn’t describe every set piece, and so the film inevitably conveys some information independent of the book.
Both mediums make the most of their capabilities, though. The book In Cold Blood familiarizes us with the Clutter family in a way that the film fails to do (most likely due to time constraints). Our greater intimacy with the Clutter family and characters like Bobby Rupp in the book serves to make the Clutters’ deaths and the impact they have on the Holcomb community more salient to us. The film is also able to stir emotional responses through its use of music, composition, and the pacing of shots. A great example of this appears in both films when Perry is hanged at the end. The audio swells, Perry breathes heavily, and the intensity increases until the trapdoor opens abruptly.
Yoder’s is one of the most accurate analysis submitted. His focus on only two terms allowed him to develop a detailed visualization of the films and books. In turn, he is able to provide specific pages in the books and times in the films when particular moments take place, in which both Dick and Perry are prominent. This gives us a strong sense of how the stories of Dick, Perry, and Capote were intertwined in real life as well as in the semi-journalistic fiction created by Capote himself.
The visualizations presented here should be considered starting points for the future of analysis by designers. The students in my class eventually applied the methods and critical skills developed in this particular analysis to develop a research & design project of their own, which was the final assignment of class.
There is much to be said about each of the analyses covered here, as well as others. For instance, we can see how by using analytic tools with quantitative capabilities we are able to dissect the grammar of visual and textual languages to better understand their intersections. All of the projects in class, including those not featured here did show a decent balance in quantitative and qualitative analysis. The challenge for a few of the students as is evident in some of the examples part of this entry is to push further for accuracy in their results. This is a skill students keep working on as they move on with their studies.
Figure 1: four shots from around a third into the film. Left is original edit, Right is chronological edit
During the Fall of 2013, I analyzed Pulp Fiction with my students in my Video Art Class for the School of Visual Arts at Penn State. One of their assignments was to produce a video and then re-edit it to tell the same story but in different order, and therefore explore how aesthetics play a role in experiencing a narrative. We went over a few examples that would give them ideas, some of the links I provided as resources included Pulp Fiction and Memento. I share them below:
Infographic of Pulp Fiction in Chronological order:
What Watching ‘Memento’ in Chronological Order Can Teach About Story Structure:
Timeline for Memento:
Memento Chronological Order (make sure to watch the original film before viewing this):
We also viewed a chronological version of Pulp Fiction which was available on line but, unfortunately, was taken down. And I also presented in class a two column set of still frames of the two versions of the film (figure 2) of the way it was edited by Tarantino (left), and the chronological order version (right). We discussed how the film has a particular open-endedness due to the fact that its beginning and ending appear to be the middle of the story. This is fairly well known but it becomes more than evident in the two column visualization I provide below that the editing of the film is not as simple:
Figure 2: Pulp Fiction, original on left, chronological on right.
To generalize that the ending is the middle is a misconception, once we take a careful look at the visualization above, because we can notice that both versions of the film are actually almost in the same order for the first 10 minutes more or less. Notice that the early crime scene in the apartment happens almost at the same time in both versions. It is when Wallace’s wife is introduced that real changes can be noticed. But this is a bit difficult to grasp because the two opening scenes for the original and chronological versions provide diverging intertextual framings to engage with the characters as they are introduced throughout the film: to view a scene at a coffee shop and a scene of a boy and a Captain offer contrasting contexts for the next scene of Jules and Vincent.
One can argue that these scenes are different because in the chronological version the scene with Captain Koons is taken out of the original order in which the film was edited. This is important to note because aside from this scene, all that Tarantino appears to have done when he edited the film to have the middle as its beginning and ending is to develop a conventional story that is told chronologically. It is Captain Koons’s scene, then, that appears to stand out in the chronological order, because it functions as a flashback of Butch’s early childhood–notice that in the original version of the film the very next scene is Butch waking up from a dream before his boxing match (hence linking him to the flashback).
To stay true to a chronological timeline the scene with Koons has to be moved to the beginning of the film because this moment happened much earlier than all the other events. And this makes Koon’s scene the only exception to an otherwise minimal shift in the middle becoming both the beginning and ending of the film. But the flashback could be considered part of the chronological order because it is really Butch who is reliving something in his dream, and this reliving can also be considered part of his present. This then complicates the basic premise of the middle of the film being the ending and beginning of an otherwise chronological edit. What follows shows how the film would appear if we took the now conventional notion that the middle is the beginning and end:
Figure 3: Pulp Fiction, original on left, chronological on right. Sequences visualized to trace the chronological ending of the story.
When we look at this visualization we notice that the chronological ending of the film does not fall in the middle of the original film as is commonly argued. This means that the story is further edited. When breaking it down in more detail with color-codes, we can notice the following differences:
Figure 4: Pulp Fiction, original on left, chronological on right. Actual editing of chronological sequences for both original and chronological versions.
We can see that the original film does not fully follow a chronological order that was simply edited to make the middle of the chronology the beginning and ending. The chronological ending of the film takes place about two thirds of the way into the film, while the ending of the original does fall more or less around the middle of the chronological version. But even when this happens we can notice that parts of the chronology are moved around to enhance the experience of the story. For instance, the opening of the original takes place just before we reach the middle of the film, meaning that this part of the story is part of the ending, of course. We can look at each of the other segments and notice have they are shifted to tell the story in a way that will be more interesting than it being simply chronological.
If we number the order of the original edit and juxtapose it with the chronological version, we get this:
1 | 4
2 | 2
3 | 6
4 | 1 & 6
5 | 3
6 | 5
I numbered 1- 6 the chronological sequences of the original version (left column), and repositioned them into the chronological version of the film (right column). We can notice that the two opening scenes are different (diner and Captain Koons), but the very next scene is the same (Jules and Vincent in the car on their way to do a job). The second sequence in the original version is then split in order to turn it into the final chronological sequence (6, Jules and Vincent finishing the job to end up back in the diner), this is why 3 and 6 match about a third of the way into both film versions. Notice that six then comes together with sequence 1 to end at the very middle of the chronological version and match sequence 4 (Vincent and Mrs Wallace) in the original edit. It is sequence 5 (Butch fighting, escaping, running into Vincent, and Mr. Wallace, confronting the gimp, and the eventual get away) that is the actual chronological ending of the story, but we see that in the original edit this one is followed by section 6, which is the scene of Jules and Vincent at the diner, this is also sequence 1, as we know.
So, to say that Tarantino merely took the chronological development of the story and split it for the middle to be the beginning and ending is really not correct–this is what appears to be commonly assumed by some people when they think of the middle of the story being the ending and beginning of the actual film. There is much editing at play which makes this film more complex formally speaking.
But the editing is not so radical because going back and forth between closely related timelines is quite common in films. What is peculiar of Pulp Fiction and some other films by Tarantino is that they are edited as though things are happening now, there is no clear hint for the audience to acknowledge that we are going back and forth in time. Viewers must acknowledge this as they try to make sense of the story. This approach by Tarantino challenges the cinematic aspect, because the audience must remind themselves that they are viewing a film and they must make sense of its sequences. The audience must try to make them fit so that the story comes together; but this must take place in the viewers’ minds. Pulp Fiction, arguably, is a reflective exploration of how we come to engage with films, how the process of editing can become a form of communication that also questions how we try to make meaning of the content being experienced.
Another challenge of the film is the cultural stereotypes it presents, and how viewers must question them as well. This aspect of the film is more open ended, and Tarantino has been accused of promoting certain stereotypes, particularly of African Americans. Those questions are very important to discuss, and must be. And I take the time to discuss them in actual class. For this post, however, I focused on the formal aspects of Pulp Fiction.