For a designer, visualizing data always presents certain difficulties of representation, bias, and legibility. However, when working with the digital humanities, as the data becomes more subjective, these difficulties becoming increasingly challenging. This summer I am working on data visualization for the Early Novels Database, an archive of 18th Century novels that focuses on providing in-depth, copy-specific metadata. My job requires me to use the data from the archive to create visually appealing posters and images that can be used for a variety of purposes. However, while the database contains a large amount of controlled, standardized terms that allow for a highly quantifiable look at the information, the nature of the database also requires more subjective, non-standardized cataloger notes. This is excellent for the researcher, as it provides an incredible depth of knowledge, however it makes drawing concrete conclusions from it difficult.
As a designer I have to decide whether to represent concrete data, such as statistics in the form of charts, or more subjective data that leaves the viewer to make their own interpretations. Attempting to create posters that reflect the purpose of END has given me an understanding the difficult task of representing the individuality of each novel when this information is abstracted in a data visualization of the larger collection of records. Within the desire to crystallize a message there is an aspiration to eliminate unnecessary clutter, to simplify and streamline. However, the very act of simplifying END’s data in some sense contradicts the project’s mission: providing fully fleshed-out renderings where other archives give only skeletal accounts, tagging and recording details that often get lost or go unrecorded in full-text databases and traditional catalogs. How can the researcher make claims about trends in genre, printing, authorship or narrative form if they do not have a full understanding of what is present in the novels they are hoping to represent? As I attempt to create graphic representations of the database, how can I connect these visualizations back to the novels themselves in such a way that the material objects behind the metadata are historically contextualized?
One way to do this is by illustrating the process by which a cataloger generates metadata from the pages of an 18th century novel. I’m currently drafting an info-graphic that attempts to do both: communicate information about make-up of century novels and about END as a project devoted to collecting and disseminating metadata. To do this, I’m compiling 150 images of early novels title page to generate a heat-map of what an “average” early novel title page looks and feels like. Over the “hot-zones” on this heat map, I’m including information about END––about the project, how the data is collected, and our data might be used.. I began the visualization when I started experimenting with the hundreds of title page images we have digitized. The concept that I eventually decided on was to overlay these images at a low transparency, so that the sections of the title pages with the most text are very dark, with gradations of scale showing the areas of greater or less frequent usage. Through this device, the viewer can see not only where the majority of information lies on a title page, but also what types of information occur where. On top of this is my mock title page that contains conglomeration of the most frequent types of information and representations to occur on title pages, edited to reflect information about END.
If we look at the overlay image of the different title pages, we can see different trends of where data exists and how it was presented, for instance the printer information is almost always at the bottom of the page in a smaller font and that we frequently have large paragraphs of summaries or epigraphs. This allows the unfamiliar viewer to observe differences between 18th century and 21st century graphic design. In contrast to our modern title pages, printers tended to include quite lot of information crammed onto the title page, with text, often lengthy novel-titles, frequently taking up the entire page. I also chose to include an object that frequently appears on title pages: a printer’s ornament, which is typically a small design used to fill space. Because the poster was describing END I wanted to use a design that contained a book, so I took the design from a more modern book. However, if we look at the title page for Amelia, we can see a similar ornament.
The signature on the top right corner is meant to serve as both an homage to marginalia and also a statement about the ways that novels were handled and marked. It is not uncommon to find not just names but also doodles, stamps, subscription lists, bookplates, and notes on the fly-leafs or title page of the novels declaring ownership or usage. These inscriptions are sometimes written in beautiful, almost unintelligible script and sometimes there are just scribbles on various pages. We have found intricate drawings of characters or themes of the novels as well as bookmarks and tiny, cut out figures. The purpose of including an allusion to these elements is to alert the viewer that though we see these books as precious and rare, the way that they were marked up and handled reminds catalogers and scholars that some of these books were the equivalents of our cheap paperbacks. These various details connect the viewer to the novels in a way that is not always recognized when looking solely at the database entries themselves.
This design is ideal for crossing the divide between the physical and the digital, a continual struggle for the DH cataloger, because it takes the physical object and digitally manipulates it in order to create new information and understandings. For instance, if you look at the images of title pages below, you can see much of the information that END documents. The very words on the page and the way they were put there (i.e. letter-press or handwritten), the choice of epigraph, author information, material wear, marginalia and inscriptions, etc. are all significant indicators of meaning and materiality to the researcher. In addition, the idea of using a background of title pages forces the viewer to consider their meaning, if not necessarily their original context. For example, looking at the draft of the poster, we can see not only 18th century owners’ inscriptions through the overlay, but also contemporary library markings, such as the stamp of the University of Pennsylvania library in the bottom right corner. This gives the viewer a sense of the long provenance of the book – not just the inscriptions of its first readers, but also the more standardized marks indicating the repository in which it currently lives.
The in-depth, extensive metadata provided by END’s catalogers allows researchers access to information that is rarely provided in a library catalog or large archive and thus it is data that they may not otherwise have. END’s records do not replace a facsimile of or a full-text novel but instead giving greater insight and available information by providing not only a standardized set of data, where available, for each novel, but by also recording the quirks and oddities of each individual copy. In addition to representing these unique copies, I am also interested in visualizing the kinds of histories, information, and narratives that are only possible with a large data set. I am interested in these types of information because while the data is interesting in its own right, such large data sets are unintelligible without a clear, concise visualization. The process of thinking through and coming up with that visualization is just as important and necessary as the data collection itself. Thus, for the rest of the summer after I have finished this poster, I will be working on creating one infographic poster that will hopefully tell a story about the meaning and representation of gender in 18th Century novels.