Reflective Blog Post – Mining the Dispatch

“Mining the Dispatch” is a work of digital scholarship authored by Robert K. Nelson, Director of the Digital Scholarship Lab at the University of Richmond, to increase the general knowledge of life in Richmond Virginia during the Civil War period. Nelson mentions the great mystery surrounding the city that persists more than 150 years later especially concerning what daily life looked like for a citizen of Richmond impacted by the various social, political, and economic impacts of a war. These impacts were greatly strengthened by the fact that Richmond was the home base of the Confederate army, and therefore an integral topic of study to gain a deeper understanding of the socio-political context of the era, making the lack of information on the subject all the more mysterious. This project uses topic modeling software to create graphs comparing the trends in topics and words used in articles over this period in Richmond, which when combined can provide insights into popular topics, concerns, and themes of the age.

To “break down the black box”, that is to state the sources, processes, and presentation of “Mining the Dispatch” I mainly searched through the “About” and “Introduction” sections which revealed the main source of this project to be the Richmond Daily Dispatch. This extensive digital archive was created by scholars, librarians, and technologists at the University of Richmond and Tufts University’s Perseus Project and covers topics in the city ranging from the time of Lincoln’s election in 1860 through the evacuation of the city in 1865. As a topic modeling program, the MALLET software package served as the main mode of text analysis for the project, being used to “uncover categories and discover patterns in and among texts”, allowing for the collection and organization of information from a distance, focusing on the patterns in content rather than the exact information presented. Through the use of this software package, not only were different topics identified, but by grouping similar words into topics, the program can inform readers what proportions and of which topics information within a given Daily Dispatch article comes from. Furthermore, I noticed that rather than providing great detail of individual data points on each graph, the project leans into the strengths of topic modeling and focuses more on overlaying different graphs and tweaking certain variables to compare trends. After breaking this project down, I can imagine that this resource can greatly aid the study of slavery in America, especially in this setting. The author himself stated that there are questions that can be derived from this data, but more traditional methods must be used to answer them. I want to know more about what avenues this sort of information can lead a researcher down. One way I find this information useful is by comparing trends in different data to look for possible correlations. For example, the picture on the home page of the site compares the trends in “Fugitive Slave Ads” and “For Hire and Wanted Ads”.

Questions from class:

What are the components? (i.e essay, interactive map, etc)
-This website provides graphs for 40 different topics, each section with two graph options showing the trends from 1860-1865. One of the graphs shows a count of the number of articles on this topic, and the other topic proportion/% print space. The most frequent words among articles on the topic and a few of the articles themselves are also provided.

What kinds of data are being used? Is the data available for broader use? Would you want it to be?
-The data being used is derived from archival data from the Richmond Daily Dispatch which can be found online, but while the source material is accessible, it isn’t available but not in a very meaningful way, at least in terms of being able to make comparisons without a little extra work. I don’t think that having the raw data available would be particularly helpful depending on whether or not you’re looking to manipulate the data for yourself.