Recently, I learned of Riot Acts by Patrick Hoehne, a PhD candidate at the University of Nebraska-Lincoln.
I’m always eager to highlight digital history by graduate students, and Patrick’s work here is impressive: compiling a database of over 2,200 instances of extralegal violence between US independence and the end of the Civil War. There’s a lot here: two interactive maps that allows you to either spatially and temporally visualize instances of violence, or query the data spatially. Not only that, but the project incudes a network analysis of types of violence and who instances of violence was directed against.
Patrick also makes his data available for use by others, not only through a table filled with notes and descriptions but also provides a simple download of the database. I deeply appreciate the culture in digital history of sharing our data, which I see as an effort not only allowing others build on our digital work but also democratizing access to our historical sources.
Another fun section of the site, iaRiot, generates fictional acts of violence using OpenAI’s Generative Pre-Trained Transformer 2 (GPT-2) and Max Woolf’s gpt-2-simple. The results generated include things like:
August: Jared Smith and five men charged with riot and assault and battery; five men were convicted in August Sessions of riot and assault and battery.
December: Group of African Americans attacked by some Yankees. Mayor says nothing about it.
Rioters burned a brothel, gutted property and allegedly destroyed an unincorporated community.
I’d love to hear more in-depth about what Patrick hopes such generated content can provide historians. He hints at this, asking what kind of patterns the algorithms are mimicking or how the language of violence is applied to fictional examples. Yet for me there’s a question about usefulness here: what insights might historians consider when using such technology? What do we hope to demonstrate, not just to our discipline but to wider audiences? I know this isn’t the core intent of the project, but it’s inclusion here raises lots of interesting questions.
Finally, Patrick contextualizes the data and history through stories — an essential component, I would argue, for any digital history project. Data and visualizations cannot stand on their own, and these kinds of dives into the history helps situate all of the digital work around the core historiographical questions and issues this work raises.
A fantastic project that I’ll be keeping an eye on as updates come.
The site is built upon ESRI StoryMaps and uses scrollytelling well in exploring change over time. But what I think is particularly effective is the design of the maps themselves. As I once wrote elsewhere, one of the challenges in visualizing geographic data over time are changing boundaries — new states, counties, parishes, census designations, and so forth make it hard to compare places against each other over time. Especially, for example, in choropleth maps — shading datapoints by county, even with a tight chronology, is difficult in the United States for the simple reason that counties get larger as you move westward. This isn’t to say that choropleths can’t be useful or done well, but it is an important design consideration.
The companion to Southern Journey gets around this by aggregating their data in hexagons (the bestagons). The state boundaries and dispossessed Native land are still there for reference. But in trying to visualize millions of people over time, the small grid of hexagons and a diverging color scale take on the primary visual load of showing change. And, like I noted above with Patrick’s work, the maps alone don’t speak for themselves — the project is deeply interwoven with a rich historical literature behind it, giving the authors a chance to not only visualize these population changes over two hundred years of history but to also explain why we see what we see.
Right now most of my time has been orienting around my new role at the Roy Rosenzweig Center for History and New Media. I’ve already written about how much it means to me to be joining RRCHNM, so instead I’ll note a few things I’m starting to work on.
The primary thing that’s getting my attention right now is Religious Ecologies, a project led by Lincoln Mullen and John Turner that seeks a new kind of synthesis in our understanding of American religious history. It’s an ambitious project with two key underlying goals: digitize the never-before-digitized Census of Religious Bodies held at the National Archives, and transcribe the data contained in the schedules that we can then use for new kinds of analysis and visualization. We’ll have some new things coming out before too long, so stay tuned. And keep your fingers crossed that NARA reopens so digitization can resume.
I’m also beginning work on the London Bills of Mortality project led by Jessica Otis. We’re digitizing and transcribing mortality bills, weekly and annual broadsides on local mortality statistics distributed around London during the seventeenth and eighteenth centuries. These bills were particularly important in plague outbreaks, and these bills represent a rich set of data on epidemiological and demographic data about early modern London. We’re anticipating an interactive database of data contained in the bills along with interactive visualizations of the information. We’re just getting off the ground with the project, so watch your social media for updates.
Finally, I have a goal in mind to ship off my finalized book manuscript to my publisher by the end of the year. Dear readers, hold me to this.
- Eric Rodenbeck: “Getting started in dataviz is an activity that’s truly moved out of the constraints of computer science and well into the mainstream of human culture, where it belongs.”
- Moviemaps from the Leventhal Map Center can link up a video discussion with automatic panning and zooming of a digital map. Seems like a great narrative and classroom tool.
- It’s not just domestication that changed animal behaviors. Proximity to humans can radically alter behavior.
- Carbon removal is a distraction to the essential priority of this decade: slashing carbon emissions.