Data

Input

The data within EarlyPrint consists of multiple components: digital images of texts, transcriptions of texts, and metadata – data that describes other data – relating to each text. The metadata for each text includes author(s), dates published, printers, publishers, booksellers, geographical locations, content tags, and linguistical information. The foundational data originates from the Text Creation Partnership (TCP), a multi-institutional project that pools library resources from around the world to create “create standardized, accurate, and faithful XML/SGML-encoded” transcriptions of early printed books. Moreover, EarlyPrint’s transcriptions and images are based upon TCP-supported archives such as the Early English Books Online (EEBO), EighteenthCentury Collections Online (ECCO), and Evans-TCP collection.

Beyond this foundational data, EarlyPrint also receives crowdsourced contributions from the public who can register on the site to become fellow digital curators. Furthermore, scholars, students, and everyday readers provide corrections and suggestions to EarlyPrint’s transcriptions and metadata to improve the reliability of the corpus data.

Output

The result of the combined foundational TCP data and public contributions is an archive with a higher confidence in transcription accuracy and improved data searchability. In addition to increased reader accessibility, EarlyPrint allows researchers to conduct historical quantitative analyses of the early English print record. For example, researchers can use an embedded N-gram viewer to search by relative keyword frequency by year. An example of an N-gram viewer searching for the word “betwixt” between 1480 and 1700 is shown below.

All of the data within EarlyPrint is openly accessible and free to download; thus, EarlyPrint’s use of linked open data, metadata, and emphasis on collaboration align with the best practices of DH.