Women of the OSS and SOE
A digital humanities project
My Process
Extant Archives and Archival Constraints
The CIA began declassifying OSS documents and moving them to the National Archives in the 1980s. More documents have continued to be released since then, so the collection continues to grow. Declassified SOE records can be found at The National Archives of the United Kingdom. Unfortunately, roughly only 15% of official SOE documents survive following the destruction of files in Singapore in 1942 and subsequent destruction in 1945, as well as a fire at the SOE headquarters in 1945.
​
Aside from missing documents, misfiled documents, and incomplete records, other archival difficulties include unconfirmed code names, records spread through various global archives, and undigitized files.
While I was able to access some primary sources that have been digitized, much of my research was pulled from secondary sources. I was unable to verify some specific information about Vera Atkins but was able to showcase those inconsistencies through the narrative maps and timelines. ​Building maps show how widespread some of the women's activity was.
Narrative Building
Building the narratives through maps and timelines bring the stories to life in a way that words can't capture. Maps showcase how the agents traveled and where they spent most of their time as part of the SOE or OSS. The timelines show how contradictory the agent's experiences could often be. In Virginia's case, the timeline shows how fast-paced but also how slow-moving her work was. When she had to move, it was of the utmost importance that she moved as quickly as possible, but it could also take months to complete a mission.
​
I started with two women because each required in-depth research and my time was limited. These women were featured in multiple books and both had their own, so I was able to gather as much information as possible to build complete profiles.
Timelines
The timelines were created using Knight Lab’s free Timeline creator. I input important dates and information into Knight Lab’s provided spreadsheet format and uploaded it into the program. Knight Lab’s Timeline creator allows for a small amount of editing for font and color, but I stuck with the basic black and white theme.
Maps
To build the site's narratives maps, I used Esri ArcGIS StoryMaps.
​
I first collected important locations in a spreadsheet based on documents and books I read. I listed the city and country, a date associated with the location, a specific event that happened there, and if the location had an important association or label, ie safe house, command center, etc. Once that list was compiled, I searched for and added each longitude and latitude. I uploaded that spreadsheet as a CSV into an Esri ArcGIS map where I added color coding and edited the pop-up boxes. Once the map was finalized, I created a StoryMap that takes the user through the map in chronological order. To make the move less disorienting, certain words are highlighted to allow the user to travel around at more specific locations.
​
I also used Google Maps to create one map of OSS headquarters to showcase another method that could be used. To create that map, I created a CSV with headquarter location name, longitude, and latitude.
Connections
The interactive network under Connections shows how intertwined espionage networks are. An interactive network provides a clear way to display the vastness and interconnectedness of the networks. The plan is to build a network along the lines of Linked Jazz but most of the people I spoke with about Gephi, understood the program from a static standpoint. Though the Linked Jazz project used Gephi, there is a more complex program used to build the interactive. So instead, as a first step, I decided to take a simpler route.
​
First, I created two CSV spreadsheets: nodes and edges. My nodes consisted of the person's name, the location where they met Vera Atkins or Virginia Hall (the two women I showcase), their title, their role alongside Vera or Virginia, and any special notes about the person. The edges were the connections.
​
Avery Blankenship, an English Ph.D. candidate at Northeastern, directed me to The Programming Historian and its network data tutorial using Python​. I used this tutorial to use NetworkX and created the graph. Printing the graph information kept giving me more nodes than I had, and I couldn't figure out why, but I continued with the project anyway. I skipped all of the modularities to see how the graph would turn out without them, and I exported my data into Gephi’s GEXF format.
​
Once I opened the data in Gephi, I opened a data table and could see what nodes were being added to my graph. It took me a while to figure out that some of my names had spaces at the end, which created entirely new nodes. I cleaned up my spreadsheets and repeated the process to correct the problem.
​
I color-coded my nodes based on the role because I felt that it made the graph easier to explore. I went with the Fruchterman Reingold layout with label adjust.
​
When doing research into making interactives with Gephi, I found that I should install the SigmaJS exporter plugin. Using a tutorial by Volodymyr Miz, I was able to export my visualization and publish it to GitHub. This allowed me to embed the final interactive on my site.
Future Steps
This project has the capacity to grow and include as many women of the OSS and SOE as exist in the archives. For now, it's an example of how to present archival information.​
​
Research
Hundreds of women were involved with the OSS and OSE. This project is specifically geared towards those who served or traveled in the European and Pacific theaters during the war. Some of these women are even mentioned in the network connections, but with more time and research, their stories can be fleshed out like that of Vera Atkins and Virginia Hall. This additional research would come from other secondary sources, in-person archival research, and newspapers.
​
Maps
For every woman added in the future, a StoryMap would also be created. The website could also grow to include maps of SOE offices, training centers, hideouts, and sabotage sites.
​
Connections
Right now, the network is completely symmetrical. The next step would be changing node sizes to match the connection weight. This would make both Virginia and Vera's nodes larger. Additionally, the visual's Legend and node menus should be edited. The nodes feature some unnecessary information that may be distracting for visitors, so that should be cleared up. The color code for the nodes is above the visualization for now, but the legend could eventually feature the color information.
​