The NULab project that I’m working on right now involves documenting connections between newspapers in the nineteenth-century United States. So far, my work has been researching the history of each individual newspaper. It’s been an enlightening and entertaining process. (If you’re interested in one of the most entertaining stories I discovered, check out my Omeka exhibit for my digital humanities class.)
We’re pulling data from the Chronicling America website at the Library of Congress. The newspapers we have range from 1836 to 1860. We don’t have all the newspapers from that range, though. We’re adding new papers all the time. The data I’m working with right now is from the first batch of data.
One of the difficulties I encountered early into the process of research was the astonishing number of name changes each paper went through during its lifetime. To get a better sense of how many times and how often these name changes occurred, I decided to plot the changes on a timeline.
Based on a suggestion from Chuck Rybak, and armed with an excellent basic tutorial by Brian Croxall, I built my timeline using MIT Simile Timeline. I found data entry very easy using the timeline interface. (I don’t think the CSS is particularly beautiful; as I have time, I may try to make it nicer-looking.)
For many of the newspapers, the exact date of some changes is uncertain. Various sources disagree on dates, and some information is just not out there for me to find. To compensate for that difficulty, I marked each uncertain date as starting on January 1 of the year and noted the uncertainty in the comment.
The current timeline has another drawback: it can’t filter by newspaper. For some newspapers, it’s easy to tell which names are connected (for instance, it’s pretty easy to tell that the Sunbury American is connected to the Sunbury American and Shamokin Journal). It’s not so easy to see the link between papers such as the Salt River Journal, The Radical, and The Democratic Banner. I’d like to be able to filter the timeline so these connections are self-evident without reading the comments.
Here’s the final result, though as newspapers get added to our dataset, they’ll get added here too. You’ll notice, too, that my timeline spans more than 1836-1860. Though many of the newspapers exist past 1860, I decided to stop my investigations there because it wouldn’t be that helpful to our project. However, I decided to trace each paper back to its origin, if possible, as a way to get at the characteristics of the paper. For that reason, the start date for some newspapers is well before 1836.
If anyone has suggestions for how to improve this timeline, I would welcome them! Please leave me a comment.