Geographies of the Text: Mapping and Annotation
Created by Karah and Nicole. Last Updated May 3, 2017.
Maps are everywhere. They are plastered on classroom walls, displayed at kiosks in the mall, and give us directions through apps on our mobile devices. We consume maps on a daily basis, and they are wonderful tools to help us get where we want to go, study world geography, and even comprehend voting patterns in an election. But maps are much more complicated than they first appear. Often, the decisions that go into creating the maps are invisible as consumers of these documents treat maps as factual representations of reality. Todd Presner, in his book HyperCities Thick Mapping in the Digital Humanities, complicates how we perceive maps by stressing the fact that maps are products of choices rather than apolitical objects. He claims that “mapping is not a one-time thing, and maps are not stable objects that reference, reflect, or correspond to an external reality...Maps are visual arguments and stories; they make claims and harbor ideals, hopes, desires, biases, prejudices, and violences” (15). Understanding that maps (and indeed all types of data visualizations) are cultural productions that are subject to the cartographer’s choices and opinions is crucial for integrating mapping into the humanities. For the purposes of this project, the question then becomes: do maps offer any insights into understanding the meaning of the comics? Or vice versa—can the comics unveil deeper understandings of the maps? A closer examination of each of these reveals that the maps and the comics can certainly stand alone as individual rhetorical texts, but that when combined they are able to expose a nuanced understanding of the interlacing themes of temporality and spatial location, which in turn illuminates how historic patterns of unequal power distribution has led to the marginalization of both people and places.
Connecting Mapping to Digital Humanities
Before diving into the rhetoric produced by the maps, it is important to recognize that maps, like literature, operate as a form of text. According to J.B. Harley, “Maps are text in the same senses that other nonverbal sign systems—paintings, prints, theater, films, television, music—are texts” (36). He claims that maps “exhibit a textual function in the world” because they, like books are, “subject to bibliographical control, interpretation, and historical analysis” (36). Approaching maps in this light places them on an equal footing with literature because both are inherently rhetorical as a result of being “the products of individual minds” that are influenced by “the wider cultural values in particular societies” (36.) Just as a close reading can be applied to novels to reveal themes and criticisms about the social world, a similar level of detailed analysis can also be used to approach maps in order to reveal information that is hidden beneath the surface.
The maps and timelines we will be focusing on throughout this essay include a geospatial map plotting publication locations, a timeline to see when and where the comics in our data set were released, and a non-traditional map that plots moments of marginalization in one of our digitized comic books. These visualizations have been generated using the online geospatial mapping tools: Tableau, Neatline, and Palladio. While our main focus is to analyze these maps and their ability to contextualize the comics, part of our goal in also to narrativize the procedure we used to make these maps in order to expose the behind-the-scenes decisions that went into their production.
Introducing the Mapping Process
To begin the mapping process we reorganized the data set to include only the information relevant for mapping, such as latitude and longitude, publishing location, and copyright year. Each of the three DH tools required the data to be formatted in a slightly different way. For example, Tableau required an Excel file with the latitude and longitude in their own columns while Palladio needed a .csv file with the coordinates listed in the same column, separated by a comma. These adjustments were relatively simple manipulations, and taking the time to generate the lat-long coordinates drastically improved our final mapping visualizations. For instance, Tableau has the ability to approximate coordinates for geo-codes, country names, and city names. However, when we attempted to use this feature, all of the data points were within the United States with places like “Paris, Texas” appearing instead of “Paris, France.” This highlighted the USA-centric nature of Tableau, and necessitated the supplementation of our existing data with accurate lat-long coordinates, which we generated using an online GPS Visualizer tool.
Another challenge we faced was filtering out the excess data we had within the master data file that slowed down not only our manual retrieval of relevant mapping and timeline data, but also the ability of Tableau and Palladio to process the data input. This was easily resolved by removing unnecessary data columns and keeping only what was immediately useful. Once the data was reduced to a more manageable amount, we found that there was still a lot of null or void data where information about the comics was simply unknown. Plugging these null points into Tableau and Palladio produced slight blemishes in the final mapped products. Looking at the Palladio map for example, there is a point off the coast of Africa at zero degrees longitude and zero degrees latitude that does not refer to a specific publication location, but instead illustrates how the tool handled null data. While on one hand it seems silly to keep these null data points and have them mis-represented spatially in the middle of the ocean, we decided to do just that because keeping it meant maintaining the integrity of our data set, and thus the integrity of our translation of the data into a visualization. Removing the null locations would mean removing a large portion of our comics from the map leading to the unintentional production of silences.
Once we filtered our data, we plotted the publishing locations using Tableau and Palladio. A benefit of the Palladio map over Tableau was that it could demonstrate which publishing locations were more or less prominent producers of comic books during the early twentieth century as the size of each point reflected the volume of comics published at each location. Visualizing the data in this way provides some historical context of the comics by not only suggesting which publishers were leaders in the comic book industry at the time, but also by indicating where the comics were produced, and therefore connecting the comics to the local culture that is associated with different places. Both the Palladio and Tableau maps revealed that our comics were produced largely in Eastern United States and Western Europe highlighting the Euro- and U.S.-centric nature of the comic book industry in the twentieth century, but also demonstrating that it was by no means a monopoly as comics were popping up in scattered patterns across Latin America and Southern Africa. Furthermore, zooming into the Palladio map reveals that comic production was not confined to large cities in the same way that many other industries were at the time. For example, Racine, Wisconsin, a small town that today only consists of approximately eighteen square miles, made a large name for itself in the comic book industry publishing a higher volume of comics than Chicago, London, and Paris. Thus, the Tableau and Palladio maps indicate that while the majority of our corpus was published in big cities, large portion of them come from non-urban areas. Ultimately, the maps produced by both technologies combined allowed us to recognize that the comics included in our data set were mainly published in Western Europe and Eastern United States––both places with high levels of industrialization at the time, and both places with rich histories of imperialism, capitalism, war, and social stratification. Understanding historic patterns of the publication locations therefore has the power to further nuance the way we approach the content of the comics themselves.
One limitation of the Palladio map, however, was that it did not allow us to provide additional information about the data when the user clicked on one of the points. This failure to include depth to the visualization makes the map somewhat inaccessible, or thin, if we think of it as opposed to what Presner refers to as “thick mapping.” Thick mapping refers to the processes of collecting, aggregating, and visualizing layers of geographic or place-specific data, and is defined as the “extensibility and polyvalent ways of authoring, knowing, and making meaning,” (15). Without additional contextual information about the comics themselves (often the information valued by humanists) the Palladio map is just a representation of a bunch of numbers. This is the tradeoff of using Palladio versus Tableau: while the Tableau map does not adjust sizing of points to communicate the frequency of those coordinates in the data, it does include additional information about each comic, such as data about the author, publishing location, language, etc., when a single point is selected, thus offering a “thicker” map than Palladio. This style of mapping thus creates more meaning and context behind what is traditionally known as a simple point on a map. One complication, however, is that when there are multiple comics plotted at the same location, it is difficult to click on each point, or even determine how many are at each location.
Where the comics were published.
Using Different Digital Humanities Mapping Tools
One method to further contextualize the history of our comics is to create a timeline. Using Tableau, we generated a timeline that uses both time and place to show publication patterns in each location from 1849 to 1918. This visualization provides spatial context that allows us to link the temporal and the geographic to create a better snapshot of the setting in which our comics were produced. As the timeline reveals, the majority of our comic collection was produced in the post-industrialization era and in the midst of World War I during a peak in global nationalism. Recognizing historic events such as these, and their surrounding social implications, provides a backdrop for the comics themselves providing some explanation for the appearance of racist and sexist attitudes. However, the insight into when and where these comics were produced, is gained at the cost of losing the supplemental information about the individual comics themselves, or the thick mapping elements. One drawback to our specific timeline visualization was that the scale at the top of the chart showing the change in time is highly inconsistent. The gap between 1870 to 1876 is the exact same size as the space between 1903 to 1904, thus offering a misleading representation of change over time. As McCloud points out in his book "Understanding Comics," representing time in a two-dimensional space can be tricky, but he argues that “time can be controlled through” various strategies including “the number of panels and the closure between panels” where the closer two panels are together, the less time has passed (101). Replacing the comic term of “panels” with the phrase “increments on a scale,” we can see how our Tableau timeline distorts the depiction of time by failing to use adequate spacing to mark the difference between six years and one year. Despite this drawback to our particular visualization, the timeline serves as a reference tool for perceiving historical context, and begins to explore the intimate relationship between space and time, something that is further nuanced in the map we produced using Neatline.
For the Neatline map, we explored the idea that spatial data does not need to be confined to solely being projected on a world map, but could instead be projected on any surface including a book. By using a collage of one of our digitized comics (“Charlie Chaplin’s Comic Capers”) as the basemap for Neatline, we were able to explore the comic itself as a spatial object. This particular comic shows the film star, Charlie Chaplin, taking on various roles in brief episodic comic strips reminiscent of his short silent films. However, it also presents problematic depictions of women and African Americans. Plotting moments where female characters appear in one color and where African American characters appear in another, we were able to critically represent the comic’s marginalization of these groups of people both in terms of their overall infrequent and stereotyped depictions, as well as their often literal marginalization on the page layout. These multi-faceted layers not only offer narrativized critiques from the cartographers (the authors), but also includes embedded links to items in our digital archive and all of the Dublin Core metadata that comes with it, thus exemplifying Presner’s concept of thick mapping. Additionally, this method of mapping allowed us to restore some of what Benjamin refers to as the “aura” of a cultural object as we were able to capture the digitized book itself, complete with its wrinkles and rips, whereas our other maps failed to include this kind of detailed visual information. A limitation to this method of thick mapping, of course, would be the necessary exclusion of all nine of the other digitized comics in the archives, not to mention the thousand of comics that are included in the data set that was used in our other maps. Using this method acts as a close reading of one single text and offers a small glimpse into what the content of the comics are, how they fit into the setting under which they were created, and how the artist utilized space on the page to grant privilege to certain characters over others.
To utilize the Neatline map to it's fullest potential, view it here
To conclude, mapping both the large scale data set and the small scale comic book page allowed us to strike a balance between the costs and benefits of mapping practices in the digital humanities. In some instances, we gained a zoomed out perspective of larger patterns but lost the details of individual comics, while in others we gained access to a single comic in its entirety but lost the larger scope of patterns across the corpus. The timeline offered a visual approach to representing the period of time that the comics appeared on the market, while the maps complemented this information by representing the space and distance over which our corpus was derived. In part, this reflection was meant to enable close reading of maps as texts, illuminate the fact that there are unseen decisions that go into producing these texts, and reveal some of these decisions that went into our mapping process. In the end, our purpose of using digital maps and timelines for this project was to display the data in both an intuitive way, that would be accessible and representative to the audience, and in a critical way, keeping in mind the costs and benefits of our visualizations as cultural productions themselves as well as thinking critically about the comic book as its own spatial object capable of marginalization.