erica milde

TUNING TEXT: AN EXPLORATION IN DIGITAL MEDIA

PROJECT DESCRIPTION

Tuning Text is a browser-based data visualization for optically ‘tuning’ information channels within digitized typography. Just as a prism reveals previously unseen hues in the visible wavelength spectrum, this proof-of-concept platform allows for new interpretations and configurations of internet-native content through the mechanical matching and colorized labeling of typographical content. The platform is intended to expose knowledge-forming information within the chaos of our digital global mediascape through an experiential approach to IT-mediated interaction. 

Overview

Tuning Text is a browser-based data visualization project developed for the experiential meta-analysis of internet-native typographical information. The proof-of-concept platform utilizes text from Project Gutenberg as a mutable database of words, sentences, and themes. The data visualization removes text from the original printed format and enables the juxtaposition of key topics through the use of data-scraping code written in the programming language, Python. In the platform, key topics are assigned to a colorized spectrum that is not ultimately limited to a specific set of terms but vary infinitely. Transforming text into this colorized spectrum creates a tool for the frequency modulation of typographical information. This creates a layered view that transforms text similarly to the prism’s ability to separate light. The prism reveals individual color elements within the visible light spectrum, as Tuning Text allows for user-optimized categorization of information. This grants the viewer an ability to oscillate between the original text and a closer investigation of specific overlaps in key terms and concepts.

The interface was tested with two different texts available on Project Gutenberg. The first experiment uses the entirety of The Notebooks of Leonardo Da Vinci to isolate various subjects of study including elements of fine art, science, and geometry (Figure 1). The second experiment looks specifically at Friedrich Nietzsche’s discussion of the feminine in Beyond Good and Evil (Figure 2). In this example, sentences that contain key terms signifying “feminine” or “woman” were extracted from the complete text before searching for common key terms within those sentences. The result creates a way to explore sections containing the phrase “woman AND ____” within the content. The viewer can then “tune” in to specific concept frequencies or view the intersections between the nine subtopics. For example, viewing red text will reveal phrases containing “women and independence” while maintaining visual awareness of the overlap with the green of “women and truth”. The results provide a type of visual-muting of informational noise to expose relevant content more clearly on an illuminated digital interface. Viewing information in this way allows for a search of interconnected topics crossing between sources and mediums rather than following a linear progression from one idea to the next. 

Tuning Text is a platform concerned equally with information science and creative exploration. While existing literature and reference media is often accessible in digital platforms and archives, Tuning Text allows for deliberate fragmentation and manipulation of information as guided by input parameters.     

Statement of the Problem 

The “Curse of Dimensionality”, is an expression coined by Richard E. Bellman that refers to various phenomena that arise when analyzing and organizing data in high-dimensional spaces. Organizing and searching data often relies on detecting areas where objects form groups with similar properties; in high dimensional data, however, all objects appear to be sparse and dissimilar in many ways, which prevents common data organization strategies from being efficient (Bellman). While the curse of dimensionality refers specifically to large volume data sets that do not occur in the three-dimensional physical space of everyday experience, interface mediated interaction with large volumes of digital information is a daily exercise for many. 

 The increasing volume and subsequent fluctuation of information sources create a digital environment antithetical to the distribution of reliable educational content. Additionally, the commercialization of digital platforms intentionally alters user experience by inducing confirmation bias in order to to misinform and deny access to knowledge forming information. These setbacks to information-seeking further the anti-intellectual movement in mainstream digital culture, leaving many to react to sensationalized media outlets rather than engage in critical thinking. 

Our access to digital information is dependent on algorithms which are neither neutral nor transparent. The “ed-tech” sector approaches the inherent biases in technology by retroactively aggregating and categorizing content within new educational platforms, but a significant portion of academic content resides on outdated sites with restricted access. Platforms such as Google Scholar provide users the ability to locate peer reviewed journals and academic content but the dominance of linear, typographical communication prohibits dimensional thinkers from quickly extracting relative and comparative content. 

Significance of the Problem 

Our world consists of more than just matter and energy, so the practice of socially responsible design must also flow into the “flow of information itself.” The internet has forever changed standard models of information dissemination, creating a participatory environment that privileges immediacy and feedback loops over certainty and linearity,  Where communication was once point A-to-B, information now radiates dimensionally and instantaneously across our globally interconnected web. 

Pervasive problems, such as the encounter of the digital divide and net neutrality laws impact communication in often frighteningly unseen ways. Algorithms can both serve to perpetuate and to expose social inequalities coded directly into our daily digital experiences. In the absence of a real movement towards digital media theory, exponentially increasing content does not create more knowledge, but rather information overload. This has a tendency to disconnect the media consumer from context and ultimately impacts capacities of differentiation, key to the formation of opinions through the juxtaposition of ideas. Search engines and social media networks have come to the forefront in our mediascape, but ultimately these networks largely exploit user data to direct interaction towards monetization.

Tuning Text is designed to transform the existing system by applying user-determined, variable ‘lenses’ onto digitized, typographical, information. By extracting, overlapping, and color coding information, a non-linear approach to linear communication can be achieved, allowing users an experience of exploratory education rather than perpetuate the current trend of cognitively blind consumption.

Design Objectives

● Unite design ontology and literary content by finding a balance between non-typographic visual communication and linguistic analysis.

● Question the design maxim “Form Follows Function” by experimenting with alternative functions of typographical information. Can new forms dictate new functions? While this design guideline states that function is the driving force behind the physical form of an object, in a McLuhanesque inversion, the form of the presentation medium can alter the pre-existing function of a work by creating a feedback loop between form and function that simultaneously evolve. 

● Explore digitized literature and textual knowledge as an artistic medium; literature, as a component of multimedia collage, that can be manipulated with strategies used in intuitive visual communication. 

● Unite mechanical matching (backend programming code) with the process of imaginative making.

● Modify the methods of receiving information without intentionally manipulating the informational content: using the interactivity of digital mediums to extract and further fragment text while maintaining a relationship to the original written form.

Collectively these objectives serve to provide alternative perspectives by extracting content from the material form of a printed text to create a layered and multifaceted approach to research and information consumption. 

Project Justification 

Tuning Text was inspired by a desire to move from the passive consumption of information to the exploration of the creative potential of emergent digital communication technologies. Investigation into the philosophies and written work of Buckminster Fuller lead to Synergetics: Explorations into the Geometry of Thinking, a discussion of the relevance of frequency modulation in thought processes.  

Thinking is a putting-aside, rather than a putting-in discipline. Thinking is FM - frequency modulation- for it results in the tuning out of irrelevancies (static) as a result of definitive resolution of the exclusively tuned-in or accepted feedback messages’ patterns differentiability. And as the exploring navigator picks his channel between the look-out-detected rocks, the intellect picks its way between irrelevancies of feedback messages. (Fuller 509.10)

Fuller’s observations, which emerged well prior to the oversaturation of widespread digital information, are even more relevant to today’s mediascape. While the internet enables the convergence of type and image, information is still fundamentally based on the linear constraints of computer programming languages. Breaking free from default modes of media presentation conditions the possibility of the emergence of new perspectives and new modes of consumption. Design for digital media is typically dominated by software engineers and design professionals, relegating thought to two dimensions but ultimately the information we consume is mediated through a physical world of dimensional objects. Digital applications too often deny “physio-spatial” and other “non- neurotypical” thinkers the ability to manipulate, overlay, or fully explore the multi-dimensionality of information.

Design Research

Tuning Text emerged as an exploration of alternative, interactive visualizations within information science. Through sketching and documenting modes of transmission and dimensional qualities of information dissemination, a programmable strategy for fragmenting and restructuring data was devised. Additional testing and utilization of available mediated formats of academic content led to the analysis of strengths and weakness in existing modalities of information dissemination.  

The project is designed to be used in-conjunction-with, rather than instead-of existing platforms for searching information. Wiki-Links and Google Scholar are two highly used platforms dedicated to exploring digital reference and scholarly materials. Wiki-Links features a method for gathering massive amounts of cross-document reference data before outputting a visual representation of overlapping and related entries from Wikipedia. This mind-map presentation of Wikipedia's content provides an opportunity to explore and discover new information while enabling spontaneous learning more akin to physical libraries. Google Scholar’s web based search engine indexes scholarly articles, peer reviewed academic journals and other academic content. While these existing search modalities streamline the user’s search for relevant content, Tuning Text would follow as the secondary mode for exploring content hosted on various databases.  

Design Process    

Creating a functioning prototype required development of basic data scraping processes. I began the project with minimal coding experience and no previous knowledge of the Python language as used for data mining. Through the use of tutorials and extensive experimentation, a script was developed to complete the scraping and matching processes required for the user interface to function. Tuning Text’s backend Python script retrieves a text file from Project Gutenberg, before parsing, tagging and exporting the data into a json file for view within the online interface. While the process of mechanical matching is directed by the specifications written into the script, this data-mining code and interface design could be utilized with any digitized content. Figure 3 shows the coded script for Beyond Good and Evil, revealing the terms and subterms selected to display on the interface. Ideally, the program will be adapted into an interactive platform that allows the user to input desired content and specify terms to return. While the proof-of-concept is shown with preselected parameters and terms, ultimately the program will feature a more robust user interface that places input control with the user. 

Summary

As big data continues to inform anticipatory design, our physical and digital realities are responsively shifting, requiring long overdue ethical and moral responses within the digital sphere. As stated by Christian Rudder in Dataclysm, the illusory nature of digital content exists with an inherent dichotomy. 

It’s a delicate illusion, the Internet; imagine a carrot sliced so cleanly that the pieces stay there in place on the cutting board, still in the shape of a carrot. And while this tension—between the continuity of the human condition and the fracture of the database—can make running a website complicated, it’s also what makes the story go. 

It is within this place of tension, between the apparent and the assumed, that responses emerge and alternative modalities for IT-meditated content can be formed. While this initiative cannot change the structure and/or form of existing content, having access to Tuning Text as a browser plugin could fundamentally change the way individuals approach digitized content. The platform could enable instant comparative research and allow for reframing information by eliminating noise within digital mediums. Enabling the viewer to determine their own parameters for comparison and analysis of content will increase the internet-using population’s inclination to both question and reassess information sources.

Tuning Text was created as part of a MA in the Media Studies program at The New School in New York City. Developed independently, Tuning Text requires interdisciplinary collaboration to further the project beyond a basic level of prototype programming. Working with more expertise in computer science and programming languages would optimize performance and streamline the product, while maintaining the functional essence generated from within a framework of industrial design. The project mission is to innovate for new mediums by applying the principles of physical design to the dissemination of information through media products.  Efforts focus on finding balance in the production and consumption of information while incorporating the dimensional nature of physical reality into IT-mediated space.