With the majority of the encoding documentation completed with the exception of a few minor tweaks when new instances come to light, I transitioned to another step of data processing. When first exploring crowdsourced transcription a few years ago, we explored a variety of options before deciding on Zooniverse. While initially created with STEM classification applications in mind, recent projects and platform expansions have demonstrated its potential for humanities application. To present our application for use in humanities projects, Dr. Amy Giroux and I participated in the Building Capable Communities for Crowdsourced Transcription Institute at the University of Minnesota—an institute funded through the Institutes for Advanced Topics in the Digital Humanities grant awarded by the NEH's Office of the Digital Humanities. Not only did we share our own insights with others, but we gained insight for application by others who encountered different problems and provided relevant solutions. While Zooniverse is not without “growing pain” problems, support from the project PIs and the cohort networks are a crucial part of adapting humanities processes and methodologies to (sometimes unforgiving) digital requirements.
When first uploading to Zooniverse, I selected a sample set of letters organized according to a selection of themes. A variety of factors influenced the themes and letters organized within them. The themes—Items/Goods, Social Network, Amsterdam, and Mtgs/Travel—reflect the networks PRINT hopes to track and recreate visually.
Figure 1 shows the Themes (called “Workflows” by Zooniverse) represented as tiles
"Items/Goods" contains letters related to economic networks, specifically correspondence related to trade and the movement of goods. "Social Ntwk" contains letters related to correspondence across networks of social interaction. "Amsterdam" seeks to engage international crowdsourcing communities by uploading German language records from the Stadsarchief Amsterdam. "Mtgs/Travel" captures the institutional networks along which Friends traveled to other meetings or on missionary journeys.
With user experience in mind, letters were selected based on content and readability. Based on keyword metadata and cursory reads, I assigned themes to the letters in a new Theme column within a master spreadsheet of exported letter data. I also assigned a legibility grade (A, B, C, D) in another unique column. Each letter was assigned a theme and legibility grade, simplifying the selection process for future uploaded letter sets. Filtering by Theme and Legibility, I more closely read the letters to select 30 to 40 letters per theme. Those with an A or B grade were then evaluated based on content. Selecting from both A and B grades provides a varying level of difficulty, intended to retain user interaction as the project difficulty can scale along with user familiarity.
Figure 2 shows the Zooniverse transcription interface that appears once a user selects a letter from a theme set. Both the “Field Guide” button at the bottom of the toolbar and the “Tutorial” tab connect users to secondary educational materials developed in previous project steps. These materials are available in a variety of formats with different learning styles in mind, and are intended to act as both a FAQ and resource to aid in accurate transcription. However, participant survey data shows that users rarely access these materials. While other project teams can assess ways to address this, it remains to be seen if proposed solutions are compatible with the limitations of the platform.
Figure 2 shows the Zooniverse transcription interface
A quick glance of the interface also reveals remaining issues of complete compatibility with humanities projects. [Figure 2] While applications for STEM identification and classification were primarily simplistic per subject (or image), the line drawing format—especially when combined with cramped letters—is not entirely ideal to the level of legibility necessary to transcribe documents. While issues of complete compatibility exist, Zooniverse remains one of the most powerful tools to implement crowdsourced transcription efforts and to best leverage their contributions to further PRINT’s project goals.