OCR and Asian American History: Creating a Search Tool for the Survey of Race Relations Records

Abstract

The Survey of Race Relations records is a large collection of documents housed at the Hoover Institution from a survey conducted in the 1920s led by Stanford professor Eliot G. Mears. The collection is widely used by scholars for records of Asian American people’s lives during this period, especially as few records exist. Stanford and the Hoover Institution offer some online tools to navigate the archive, including a finding aid and an online catalog. These tools are helpful but limited in capability. Additionally, the documents are inconsistently named, and the contents of the documents are not searchable. In this project, I improve existing ways to find documents in the Survey of Race Relations records by running the documents through optical character recognition to make them full-text searchable, classifying them into categories based on location, type, and population, and creating an interface to search for and filter through documents.