They came to Stanford from 114 companies and 31 universities – with thousands more joining online – for the second annual Women in Data Science conference. Simultaneously, 78 satellite events in 26 countries were underway, on every continent but Antarctica. The conference trended nationally on Twitter all day.

Stanford statistics Professor Susan Holmes, left, takes questions after her presentation at the 2017 Women in Data Science (WiDS) Conference. At right is moderator Margot Gerritsen, senior associate dean and director of Stanford’s Institute for Computational and Mathematical Engineering. (Image credit: L.A. Cicero)

The day-long conference, held last Friday by the Institute for Computational and Mathematical Engineering (ICME), featured 16 speakers, all women, from government, academia, industry and nonprofit sectors, discussing a range of topics in data science, including the discipline’s role in artificial intelligence, national security and health care.

One goal of the conference, said ICME Director Margot Gerritsen, was to highlight the latest data science-related research and to learn how leading-edge companies are successfully using data science.

Although the conference was strictly technical – with a focus on work from some of the world’s leading data scientists – it also brought together women in a field typically dominated by men, providing a forum with which to inspire and connect female scientists and engineers. “There is this sense among women – more than men – that you need this innate ability to do data science,” Gerritsen said. “We need to change that attitude.”

Common focus

On Friday, the meeting room at Stanford’s Frances C. Arrillaga Alumni Center offered a sense of commonality. Many of the speakers touched on the necessity for women to pursue careers in data science.

“Data science is definitely an area in which women can and will excel. I’m very pleased to be invited to present. It’s vibrant. It’s inspiring,” said speaker Stephanie Gottlib-zeh, president of Agyleo Sport, a French venture capital firm that specializes in new digital technology.

“The women are here. We are not going away,” said Janet George, fellow and chief data officer at Western Digital.

Susan Holmes, a Stanford professor of statistics, began her presentation on a political note. “I decided to wear a black suit and tie because the president suggested that women ought to dress more like women,” she said, referring to unconfirmed online reports that President Trump has directed female White House staffers to “dress like women.” In a room of nearly all female scientists, her point had a particular potency.

Holmes continued with a technical presentation about the ways statistical analysis can bias large data. Holmes argued that depending on which statistical method one picks, results may vary, and scientists should be transparent about which statistical tools they are using. In one Nature study, she said she found there were 204 million ways to analyze the same data set.

Other speakers included Yael Garten, director of data science at LinkedIn, and Diane Greene, senior vice president at Google Cloud. Deborah Frincke, director of research at the National Security Agency and the conference’s afternoon keynote speaker, spoke about how the NSA executes “mission-oriented research” using big data. “If you are trying to deliver an advantage over an adversary, whoever they are, you need to be on the cutting edge of your game and certain kinds of technology,” she said.

Megan Price, executive director of the Human Rights Data Group, a San Francisco-based nonprofit that supports humanitarian rights groups with data analysis, spoke about how to accurately estimate Syrian conflict victim numbers. A death count is difficult to estimate, she said, because the same victim is often separately reported to officials two, three, four or more times – or not at all.

By looking at the similarities between death report data – location, time of death, age of victim, etc. – her team was able to get a clearer estimate of victim deaths. One must also look beyond the data. “It’s important to tell their stories, to say their names,” Price said.

Tapping a need

Stanford launched the inaugural WiDS conference in November 2015 to highlight women doing amazing technical work in data science, in both academia and industry.

“We wanted to educate and inspire more women to get involved in the field. This was a Stanford-wide initiative from the start with the Office of the President, Statistics Department, Stanford Biomedical Data Science Initiative, Mobilize Center and Computer Forum all working with ICME on the conference,” said Karen Matthys, WiDS co-director and ICME executive director, external partners.

When 6,000 people logged on to view the live stream over the course of the first conference, the organizers knew the program had tapped into a worldwide need to connect female data scientists.

“We wanted to make this year’s conference not just bigger, but available to more people in more places,” said Judy Logan, the WiDS co-director and ICME marketing program manager. Thus, the conference followed an “open-source” format, providing a live stream and a virtual check-in via video with other ambassador programs across the globe, including in such places as Madrid, Luxembourg, and Beirut.

“We’ve gone from being a one-day technical conference to a global movement of women dedicated to inspiring the next generation of data scientists.” Logan said.