New from the Library of Congress: Newspaper Navigator

Historical researchers rejoice!

We can now freely search more than 1.5  images from American newspapers dating from 1789 through 1963.   

LC Labs’ newly launched  Newspaper Navigator leverages machine learning and the substantial digitized resources of the Library’s Chronicling America website.

For decades, Chronicling America has allowed users to discover the text of newspapers in its collection using optical character recognition. But until now, images in the collection of 16 million pages were not searchable using traditional search engines.  Their discovery depended on often tedious browsing through relevant scanned articles. 

How does it work?

Once the user enters a keyword search, a gallery of images display.  Users may show results as the default gallery or as a list. They may search or sort by date (1900-1963): from oldest to newest or newest to older; enter a start and end date, and may filter geographically by state.

Clicking on an image allows you to download it, view the originating issue in Chronicling America, learn more about the newspaper title, or get the image’s URL in order to cite it. 

The Newspaper Navigator project was inspired by LOC’s Innovator in Residence, Benjamin Lee. In this video, Lee offers a demo, explanations of how the AI training, and an overall tour.

You may curate your images in a My Collection space. Your may also train the AI tools to find relevant or visually similar images by hovering over image results and labeling them with plus or minus signs.  After selecting a group of images, clicking on Train Your AI Navigator! will all you to re-query the app for results based on what it learned from your previously preferences. AI Navigators may be labeled, saved, and continually refined. 

According to the LOC’s September 15th press release:

To create Newspaper Navigator, Lee trained computer algorithms to sort through 16 million Chronicling America newspaper pages in search of photographs, illustrations, maps, cartoons, comics, headlines and advertisements. The idea for Lee’s groundbreaking project began with a Library crowdsourcing experiment by 2017 Innovator in Residence Tong Wang called Beyond Words, which invited members of the public to help identify cartoons, illustrations, photographs and advertisements in World War I-era newspapers. Users could draw boxes around visual content on a page, transcribe captions or review other users’ transcriptions.

Simple searches work best here. It’s best to begin in Chronicling America for the text content. Some of the image information appears a bit garbled, but . . .

Newspaper Navigator will be celebrated by K12 teachers and student researchers.  Imagine the richness these visuals add to History Day projects, and to biographical, historical, and cultural research in general!

Even more to celebrate: All Newspaper Navigator photos in the app and the dataset are in the public domain.