For the past several years, the DH Lab has been working on a project, TOME, aimed at visualizing the themes in a corpus of nineteenth-century newspapers. In designing this tool, our central motivation was to be able to more clearly trace the various and often conflicting conversations about slavery and its abolition that were taking place in these papers, which spanned multiple audiences and communities. (More info here).
Around the same time, a team at the University of Delaware launched the Colored Conventions Project, aimed at recovering the advocacy work performed at the Colored Conventions–organizing meetings in which Black Americans, fugitive and free, strategized about how to achieve legal, labor, and educational justice. Among the key interventions of the CCP project is to emphasize how, in the nineteenth century, organizing work took place in person as much as on the page; and how this work was performed by collectives as much as individuals.
Taking this scholarship into account, we realized that the story told in the corpus of newspapers that we’d assembled for the TOME project was, in all likelihood, a very different one from the story told through the Colored Conventions. We thought could learn more about both conversations by looking at them together. We could ask questions like: “How did themes travel from the conventions into print, or the other way around?” “Were there people or groups who played prominent roles in one venue, or the other, or both?” “What are the key differences between the conversations that took place in person vs. those that took place in print?” And, crucially: “Who are the people or groups who have not yet been recognized for their contributions, but should be?”
In Summer 2018, Arshiya Singh (BS CS ’18), advised by Dr. Klein, began to lay the groundwork for some of the models that will help to answer these questions. What follows are a series of blog posts that document our progress.
NB: Our work employs the CCP Corpus in addition to our own. In making use of that corpus, we honor the CCP’s commitment to a use of data that humanizes and acknowledges the Black people whose collective organizational histories are assembled there. Although the subjects of datasets are often reduced to abstract data points, we affirm and adhere to the CCP’s commitment to contextualizing and narrating the conditions of the people who appear as “data” and to name them when possible.
- Post 1: Data Processing
- Post 2: Creating a Network Graph
- Post 3: Exploring Network Models
- Post 4: Update I: Detecting Communities with a Co-occurrence Graph
- Post 5: Update II: Detecting Communities with node2vec