Women in Data Science Conference discusses representation, artificial intelligence and career advancement
By Michael Biggiani
The School of Data Science hosted Women in Data Science Charlottesville in a virtual event featuring several prominent speakers followed by an in-person reception aimed at highlighting the work women have been doing in the field.
The event was coordinated by Siri Russell, associate dean for diversity, equity, and inclusion at the School of Data Science, Arlyn Burgess, association dean for administration at the School of Data Science, and Danielle D’Andrea, School of Data Science marketing professional.
“It’s not a conference about being a woman,” Burgess said in an email to The Cavalier Daily. “It is a conference by women, featuring women, for everyone that recognizes that data science has many faces and voices. That was the vision at Stanford when it started, and that is why we really jumped on early to say this is a mission we believe in and are going to push for here in Charlottesville.”
The conference – which has been held annually since 2015 – also had the additional theme of inspiring future generations of women to become involved in data science.
“We don’t see as many women data scientists as we’d like, and we’re trying to change that.”
Daisy Glotzhober, data scientist
Glotzhober mentioned how she helped to kickstart a women in analytics training program at Deloitte. She hopes it will help internally develop women practitioners of the firm with data science skills to make the field more equitable and bridge the gender gap in data science.
Glotzhober — along with fellow Deloitte data scientists Lauren Moy, Anuleka Ellan Saroja, and Tanya Balsky — served on a panel on Trustworthy Artificial Intelligence. The panel discussed other ways to make data science more equitable by considering the importance of mitigating the risks around AI and developing it in an ethical manner.
The group demonstrated these claims by highlighting some examples of the consequences that occur when AI is mismanaged.
For example, Balsky mentioned an instance in which an algorithm for an emergency room was prioritizing people for treatment based on race, resulting in people of color receiving significantly slower service than white people.
Glotzhober also cited an Amazon algorithm intended to evaluate resumes to hire employees, which had a preference for men because its database of former employees featured mostly male resumes.
Addressing these challenges is not easy, as there are barriers to implementing ethical AI, such as a lack of diversity in staff teams and lack of a framework to establish more equitable methods.
“Eighty percent of these organizations that are developing AI solutions do not have any type of framework to identify ethical risks with their AI systems, let alone communicate and mediate those risks,” Saroja said.
According to the panel, while a complete overhaul of development processes towards ethical AI would be difficult, creating a roadmap and scaling progress up over time can be practical.
“Adopting a more incremental, phased, scalable solution is something that is more realistically achievable and also more palatable for our clients,” Glotzhober said.
The panel also provided advice to young women aspiring to become data scientists.
“Look for mentors and sponsors,” Saroja said. “Wherever you are in your career journey, you need to have a network of people you can lean on for professional advice, for career opportunities, and also [to expand] your skillset. You will very soon in your career realize that learning does not stop after graduation.”
Throughout the day, four other panels featured secretive surveillance technologies, applications for data science in education, data and human rights, and data science as a career, respectively.
The virtual conference culminated with a fireside chat between Russell and Alana Karen, Google search director, and University alumna.
Karen shared her own story of entering the male-dominated field through a Dartmouth internship and provided advice for women also looking to begin their careers. She especially emphasized the importance of scrutinizing feedback.
“We’re often told that feedback is a gift, which makes it sound like it’s always a good gift — it’s always the right gift, it’s always the perfect gift for you,” Karen said. “But, sometimes it’s like the socks you didn’t want from your grandmother; like sometimes it’s not a good gift and it’s not what you wanted and it doesn’t fit you.”
Karen underlined the merit of failure on the path to success. In her book “The Adventures of Women in Tech: How We Got Here and Why We Stay” — from which Russell read aloud at the event — Karen shares the value in reaching for new frontiers, especially as an underrepresented individual.
“I now anticipate failure,” Karen wrote. “I aim for a big goal, and I always have a goal in mind, but I see this plan as living.”
The conference called attention to the added pressure a woman or a person of color may feel to serve as a trailblazer. Karen said she believes it is necessary to reflect on and carefully choose the rocks — a metaphor for the things that may weigh down one from success — that one carries.
“Yes, help others, but [be] moderate,” Karen said. “Yes, listen to feedback, but also feel comfortable finding support if you think the feedback isn’t accurate.”
For Russell, the conference was motivated by a larger mission — an urgency to create an inclusive data science community.
“Diversity and equity are issues of yesterday, today, and tomorrow and we understand that they have real-life consequences,” Russell said in an email statement to the Cavalier Daily. “The School of Data Science has a mission that speaks specifically to incorporating diversity, equity, and inclusion into all of its actions. We are committed to making that mission a reality.”
