Skip to main content

Partnering with Google Arts & Culture to Research Women in Science through Machine Learning

Mary Agnes Chase sits astride a mule

Mary Agnes Chase, honorary curator of the Grass Herbarium at the United States National Herbarium, Smithsonian Institution, on an expedition to Brazil in 1929 to collect botanical specimens. Smithsonian Institution Archives, Record Unit 229, Box 20, Folder 1.

By Dr. Elizabeth A. Harmoncurator of the history of Smithsonian women in science at Smithsonian Libraries and Archives

Women have pursued some of the Smithsonian's most exciting scientific research since at least the 1870s. Yet our records from the 19th and early 20th centuries often leave out the work women have done. Recovering the stories of women in science and their important achievements poses numerous challenges. Research tools developed in collaboration with Google Arts & Culture demonstrate the possibilities of using machine learning to carry out this work.    

Today, Smithsonian collections include more than 155 million objects and specimens. They also include library volumes and archival records. With so many objects, it can be difficult to uncover information about the women who worked with and created these collections. In our collections, sometimes women were listed with their husband's name (such as Mrs. C. D. Walcott). Some women worked without pay or held titles that obscured the extent of their contributions. Photos of women have gone uncredited. To better understand women's scientific contributions in the Smithsonian's history, we recently collaborated with Google Arts & Culture to make it easier for people to use our data. This collaboration was part of our Smithsonian Open Access Initiative, launched in 2020 in partnership with Google Arts & Culture.

Mary Jane Rathbun examines crab specimens at her desk
Mary Jane Rathburn was a world's expert in the field of carcinology, or the classification of crabs. 1894. Smithsonian Institution Archives, Record Unit 7256, Box 8, Folder: 6.

Through the Smithsonian American Women's History Initiative, Because of Her Story, Smithsonian staff have identified more than 500 women who worked in the sciences at the Smithsonian. The Google Arts & Culture team developed machine learning tools that show how curators and researchers could search the enormous volume of metadata of our historical records for undiscovered examples of these women's contributions to science.

Metadata is the information that describes our collections. For a scientific specimen of a crab, metadata might include the date the crab was collected and who collected it. This quickly adds up to a massive amount of information that is too large for researchers to investigate manually. The machine learning and data visualization tools created by Google Arts & Culture illustrate new ways that curators and historians can more easily find entities like women's names, dates, and locations of work in our collection records and metadata.

One of the visualization tools is a visual map that reveals connections between people and places in the metadata.

Handwritten index card for Idotea phosphorea Harger collected at Casco Bay, Basin Cove, Tide Mill in Maine
This taxonomy card identifies two specimens collected by Rathbun and Dandridge, identified by H. Richardson. 1911. National Museum of Natural History.

Using this map, researchers connected taxonomy cards from 1911 to three women in science. Taxonomy cards help researchers sort scientific specimens. This taxonomy card mentions Mary Jane Rathbun, likely the first female curator at the Smithsonian. In 1911, Rathbun traveled to Maine and Massachusetts with scientific illustrator Serena Katherine "Violet" Dandridge. Rathbun and Dandridge led a trip to research the colors of marine animals for an upcoming museum exhibition. The card also revealed their connection with Dr. Harriet Richardson Searle. Dr. Searle was a researcher in the invertebrate zoology department. She identified some of the specimens that Rathbun and Dandridge collected on their trip.

Sepia photo of Serena Katherine Dandridge wearing a suit
Portrait of Serena Catherine Dandridge. Bedinger and Dandridge Family Papers, David M. Rubenstein Rare Book & Manuscript Library, Duke University.

Smithsonian researchers and data scientists will continue to develop tools to mine metadata for new information about the roles women have played in the sciences throughout history—contributions that have been hidden for far too long.

Explore further in Harmon's Google Arts and Culture exhibition, Recovering American Women's History.

This project was conducted using Smithsonian Open Access, which allows anyone with internet access to download, share, and reuse millions of the Smithsonian's images and image metadata. 

Related Posts

Dr. Elizabeth A. Harmon is the curator of the history of Smithsonian women in science at Smithsonian Libraries and Archives. She creates digital resources and exhibitions about women in science and ensures their legacies are ethically described and discoverable in historical collections.

Back to Top Back to Main Content