│By Vanessa Bateman, ESEH-Gale Fellow│
When I received the 2024/2025 ESEH-Gale Non-Residential Fellowship in Digital Environmental History I was just starting the early stages of my first solo book project. I had done enough research to develop my book’s main themes, structure, and research questions, but I had not started the writing process because I still had some gaps to fill.
As someone new to the Digital Humanities (DH), I applied to the Gale Fellowship because I wanted to learn how DH methodologies could elevate my research and eventually expand my output beyond a traditional academic book. As a Gale Fellow, I received training in different research and analysis methods that could be achieved in Gale Digital Scholar Lab, and access to the Gale Primary Sources.
Below I share an account of how I had to problem solve and pivot my research in the digital space, and some findings I made that will be useful in the future. Sharing moments when research didn’t go according to plan, I believe, is just as important as a polished finished project.
Visual Media as Conservation Activism
I am currently researching how visual media was used to promote wildlife conservation throughout North America in the twentieth century. Visual material I research includes photography, film, and educational programs for children. I ask: What are the potential real-life consequences on the animals and environments represented in image? For instance, in the early 1900s the Audubon Society’s magazine Bird-Lore published photos of birds killed for their plume feathers to promote the protection of species facing extinction in the United States.
Other publications, such as Outdoor Life (below), used photography to sway readers towards the protection of game-animals – shaming “sportsmen” hunting in excess or killing protected species. Photography also had the power to help establish protective legislation, such as the formation of a wildlife refuge, which I have previously written about.
Pivoting a Plan
For the ESEH-Gale Fellowship, I had proposed to focus my research on the wildlife films I was studying using the Environmental History archives and the National Geographic Virtual Library. I wanted to create a digital map of where films were exhibited over time. However, I couldn’t build a map using the Lab alone, but I could use research from the primary sources available to build a dataset that could then be used with free mapping and GIS tools available online. And – there were no hits to any of my searches.
I pivoted my plan by putting digital map-making on pause and decided to experiment with the tools the Lab offered instead. At the same time, I wanted to learn how to transform the research I had already done into a digital dataset and directly input it into the Lab.
From the Physical Archive to a Digital Dataset
Research at a physical archive is very different from a digital archive. It requires planning ahead to get the best results. Most archives require researchers to make appointments, consult with staff, and use finding aids to select which specific boxes need to be pulled out of storage.
For the Gale Fellowship, I wanted to use my research from the National Audubon Society records at the New York Public Library. I have over 1000 photos of hand-written correspondence, board meeting notes, promotional pamphlets, photographs, film and lecture scripts, audience comment cards from film screenings, drafts for legislation, and the list goes on. I made an Excel document of which boxes photos corresponded to, but I didn’t take notes of the contents to save time.
How to turn photos into a dataset – and quickly? Text extraction. I learned if I opened photos on my computer, I could select the text and copy and paste it into a dataset. This method was never 100% accurate, but neither I learned was OCR (optical character recognition technology). Luckily most of the material I had was typed and easy to extract. Handwritten material, however, was impossible to translate this way.
Two easy solutions to organising my material were to use the Upload Template from the Lab or directly input data into the Build Content tool. After a test run of 50 documents I had extracted/transcribed into the Lab, I quickly realised this was not the way to use the Lab to its best ability. My material was too specific (such as correspondence letters) and analysis was not surprising.
I learned that while I need to eventually build a dataset from all the photos from my research trips, DH methods might not be that useful to writing my book. For the remainder of my fellowship, I wanted to get the most out of my access to the Gale resources, so again I had to pivot.
Two Solutions: Expanding & Narrowing with Gale Primary Sources
How could I adapt my very focused research to a vast digital database that didn’t have the material I was looking for? I came up with two very different solutions. First, expand my search much further than I had before – to study the relationship between conservation and photography over time as reflected in the widest selection of sources. Second, to narrow my focus by using one database and focus on one author.
Expanding & Experimenting with the Lab
During my fellowship training sessions, I learned several techniques on how to get the best results when doing a search in Gale Primary Sources, most of which can be applied to other digital search engines. This included how to cross search across multiple sources, always search the “Entire Document,” and how to use the proximity function.
A proximity search is as simple as using “n” + a number to choose how close in proximity you want one word to another. My search of “photo* n6 animals OR photo* n6 wildlife,” produced some interesting results in the Term Frequency page (above) when I searched “photography” within the results. It showed how there was an increase in documents with this word at key moments: the early conservationist movement from the late nineteenth to early twentieth centuries, and the modern American environmental movement in the mid-twentieth century.
Continuing my experiment, I built several test content sets in the Lab, including searches of “Audubon Society” [with limits to publication countries of Canada and the United States after 1876] and “Conservation AND wildlife OR animal.”
The first produced 451 sources, the other over 10,000, and I used the first 10,000 to build a content set. With so many documents, the Lab’s Clean tool was very useful to make analysis and visualisations more accurate. These experiments confirmed some trends I was already working through from my previous research: how gender was distributed over time in these sources, such as how women played a role in the founding of Audubon Societies across the United States.
Narrowing: National Geographic Virtual Library
With access to the National Geographic Virtual Library, I was delighted to find articles I had never found by authors I was researching. One of them was Arthur A. Allen, an ornithologist who was also a photographer, filmmaker, pioneer in audio recording bird songs, and member of the National Audubon Society. In a 30-year time span, Allen wrote 18 articles in the popular National Geographic Magazine, which is far less than the articles he had published in scientific journals for a scientific readership.
I wanted to learn how the intersection between visual media and conservation was reflected in Allen’s writing over time. Again, I faced some challenges: the National Geographic Virtual Library functions separately to the search engine in Gale Primary Sources, and data must be manually input into the Lab. But by now, I was a pro at extracting text, and pasted OCR text directly into the Build Content tool in the Lab.
Topic Modelling revealed that the protection of birds or the environment was not the primary focus of Allen’s writing: instead, 40-60% of articles were about nesting, followed by use of new technologies in photography and audio recording. Despite not revealing information that couldn’t be learned from carefully reading the articles, DH analysis tools helped to visualise the overall themes and changes over time – and quickly.
Confidence to Use DH Methods in the Future
While my first foray into the Digital Humanities was not as smooth as I had imagined, challenges forced me to problem solve, pivot, and experiment. In the end, I learned a lot and gained the confidence to use DH tools and methods when the time is right.
This experience also made me reflect on my initial proposal to create a digital timeline map and consider why to make a map in the first place. In the future, I plan on using a tool like HistoryPin, to place historical photographs on a map or street view. This is similar to “repeat photography,” that disciplines like geography use to understand change in a landscape over time. This type of visualisation of research and change over time would be an effective and accessible way to share my research with a wider audience.
If you enjoyed reading about environmental and animal research, check out these posts:
Blog post cover image citation: Archival holdings in the American Museum of Natural History ornithology collection (left) and screenshot of a Gale Primary Sources search (right). Taken by author.
