NARA - 1950 Census
Telling the story of who we are as a nation and where we are going.
Challenge
The National Archives and Records Administration (NARA) was reaching beyond their traditional role of making records available for others to discover. With an immovable deadline of making the 1950 census available to the public on April 1, 2022 our team was tasked with building out the Transcription Tool feature that would allow website visitors to contribute to the census records making them more useful to the American Public.
The collection consists of approximately 6.5 million digitized population schedule images. NARA is using an Optical Character Recognition (ORC) Tool, Amazon Textract, to extract names from the population schedules. Given that the tool is only about 40% accurate, NARA sought to develop a Transcription Tool that enables website visitors to correct and add alternative names to the name index, in order to enhance search results.
With a tight deadline and a very public eagerness to see the records, needing the tool to work right out of the gate was crucial for success.
Contribution
Lead UX research to understand the needs and workflows for both site visitors and NARA content monitors. Validated our approach through iterative design sessions with the client and usability testing with target audiences NARA staff.
Stakeholder Interviews
Content Workshops & Task Flows
Wireframes & Prototypes in Figma
Usability Testing
Results
On April 1, 2022, NARA launched the new 1950 Census website, enabling users to access the digitized collection of population schedules for the first time.
Our team used human centered design methods paired with an iterative design process to deliver results.
The Transcription Tool successfully met the needs of both site visitors and NARA content moderators who were able to manage the thousands of initial contributions and beyond.
So far things are going great and we have had great use of the transcription tool. By 3am we had over 5,000 transcriptions already completed and by 6:15 18,634!!"
– Client Feedback
Visit 1950census.archives.gov
Workflows created to understand needs and painpoints
Wireframes and prototypes iterated on with client and tested with end users
Wireframes and prototype in Figma used for both client direction and user testing
What People Are Saying
“And though NARA has been releasing census records for decades this is the first time that NARA is simultaneously providing a first draft of the name index. We're using the AI enhanced optical character recognition or OCR tool to provide a rough draft of names we've coupled this draft with the name index with the transcription feature that we encourage you to try out this is an opportunity for you to help us refine and enhance OCR created names and make the population schedules more accessible for everyone.”
— The Archivist of the United States, David S. Ferriero
“The images being released on Friday are digitized versions of the microfilm records that have been scanned with special optical character recognition software to locate and translate handwritten names and addresses into searchable text. Mistakes are inevitable, and the National Archives is asking those who view the forms to report errors.”
— New York Times
-
Census records from 1950 being released: "This is genealogy heaven"
-
1950 Census data to be unveiled Friday, after 72 years under wraps
-
Newly released 1950 census data give public insight into family status, US history