Women in Data Science: Where We Are, Where We’re Going, and Why It Matters
By Isabella Oakes and Adila Abdulwahid, published April 2022
Historically, women have been discouraged from pursuing STEM roles. Media coverage influences girls’ lack of confidence in their own abilities. Laila Sprejer from the Alan Turing Institute notes that once women enter these roles, they face discrimination in the form of “unequal pay, slow career progression, harassment and gender bias.” The Alan Turing Institute created the Women in Data Science and AI project to create policy interventions to encourage women to enter AI careers, but there is still a large gap between men and women in Data Science and other STEM careers. In 2018, the World Economic Forum shared a report by LinkedIn estimating that women make up only 26% of workers in data and Artificial Intelligence roles across 20 leading economies in the world.
One theory to why there are not many women in STEM is that having a high-powered professional life and a rich personal life are incompatible. Another theory published in 1991 is that during college and graduate school, many female students “lose their self-esteem, are harassed by male professors and students, are excluded from crucial discussions and social interactions, and made to feel that they do not belong.” A lack of women currently in the field can perpetuate the issue and discourages younger women from entering the field. Studies have shown that women are more likely to select careers when they can identify a role model in that career path.
The article “Top Women Leaders in AI To Watch” from DataTechvibe highlights important women in the field of data science. Increasing their publicity can be a catalyst that encourages young girls to pursue similar careers. For instance, Fatmah Baothman is the first woman in the Middle East to earn a Ph.D. in Artificial Intelligence. Her research helped develop speech recognition for Arabic. Rana el Kaliouby is the founder and CEO of Affectiva, a company that helps sense consumer emotions for marketing and mental health purposes. She earned a Ph.D. from Cambridge and worked at MIT labs. Her contribution to emotional artificial intelligence allows computers to sense human emotion and understand how users feel.
Shivon Zilis is a founding member at Bloomberg Beta. She is a board member at OpenAI. OpenAI created the Generative Pre-trained Transformer 3, which is an autoregressive language model that uses deep learning to produce human-like text. Daphne Koller is a computer science professor at Stanford University and the co-founder of Coursera along with Andrew Ng. Her research involves Artificial Intelligence in the biomedical sciences. Koller started the drug development company Insitro in 2018. The list can go on, and these are just a few good examples of pioneering women in data science. Continuing to encourage women to enter these roles will provide a necessary perspective in machine learning.
Why do we need more women in Data Science?
Encouraging women to join the field of Data Science has many positives. Research is only going to be pursued if researchers have an interest in it, so having a diverse group of people increases the diversity of topics. With more diverse research comes information that will be helpful to a much more universal audience instead of research tailored to a subset of the population. According to the article, “Why Women in Data Science Are Crucial In a Data-Driven World,” 11% of data teams worldwide are 100% male. Bias is something that must always be balanced with research, so having a team that can balance their biases can lead to insights that are more meaningful. A large portion of research being done is focused on a male point of view, leaving out valuable perspectives that can help ground research to be more well-rounded and universally accessible. Products that are gender-specific and medical issues that affect women are two examples where exploring data really benefits from having multiple perspectives. An example of this is cited in the Center for Global Development’s article “Why the World Needs More Women Data Scientists,” when they reference crash test dummies being historically designed based on male bodies, resulting in safety features that were only being tested on a larger body than roughly half of the population has. According to “Why We Need Women in Data Science,” companies benefit from having gender diversity. They are more likely to exceed national financial medians, have increased revenue, and outperform teams that have less diversity.
Data Science Education
It’s clear that having diversity within Data Science is positive, so the next step is making it an accessible field. Educational opportunities that encourage applicants with different backgrounds are important to allow access to new fields. Degrees that want extensive previous experience and only strong computer science, math, analytical and engineering backgrounds can filter out many applicants, particularly women, who may have avoided those fields previously due to them being historically male-dominated fields. It’s harder to get relevant experience when positions are difficult to break into. Higher education programs need to encourage more diversity. According to BCG, while women make up around 55% of university graduates, they only account for roughly a third of graduates with STEM degrees, of which only around 66% continue onto a STEM career. This translates to only around 15-22% of data scientists being women. If the number of women graduates in STEM fields increases, this will also lead to more women in fields like Data Science.
Women in Data Science
While universities are key components in preparing women for STEM careers, encouraging women to pursue STEM fields can start very early on. Providing information about Data Science and related fields starting early on in education, even elementary school, helps educate future generations about the possibilities that girls can look up to that may not usually be offered to them. One organization working to fill these gaps is WiDS (Women in Data Science). They share many resources encouraging and supporting women in Data Science, including secondary school outreach programs. They also host an annual Datathon that puts teams on task solving real-world problems. The yearly Conference hosts many speakers from universities and companies talking about a variety of topics related to women in data science. They also provide free Workshops to help educate anyone who would like to join.
Careers in Data Science
According to the article “The Future of Women in Data Science,” companies have a significant responsibility when shaping the industry and need to provide equitable working environments. Toxicity in jobs is a huge reason women may avoid or quit a job in Data Science or related industries. Providing training with expected behavior and having ways to report inappropriate behavior without negative consequences makes it a safe working environment for women and minorities. Companies also need to provide equal pay and work to close the wage gap. No one wants to work for a company making less than their coworkers while being treated poorly, and those are major reasons women change industries.
The good news is that the number of women in Data Science is increasing. Online graduate programs, such as the University of San Diego’s Applied Data Science, allow the flexibility that can provide the educational background women need to get into Data Science positions. Encouraging women to join these programs and other STEM programs is one key to making Data Science a more diverse industry. Hiring women to be in leadership positions is critical for diversity to grow. From there, making sure that positions are flexible enough to encourage people from all backgrounds to apply and balance their home life and work allows for a much wider range of people to have successful careers. Full and partial work-from-home schedules is one example, as well as overall flexibility. With increased education options and career openings, the trend in increased women in Data Science will continue, and the future will be better for it.