Women in Data Science Worldwide

Ya Xu manages LinkedIn’s global team of data scientists that manage data science projects across the company’s products, sales, marketing, economics, infrastructure, and operations. She says the company takes active responsibility over the data they collect to ensure fairness and protect privacy. They are very proactive about how they maintain their members’ trust, either with how they share the data externally or leverage the data to create opportunities.

LinkedIn’s fairness mission is that two people with equal talent should have an equal shot at opportunities. To reinforce this, they constantly test new products to determine if a new feature introduces any unintended consequences that might impact fairness. For example, LinkedIn’s referral button allows a job applicant to see if someone they know works in the company and ask for a referral. The unintended consequence is this feature will benefit individuals who have a big network vs. the general population. She says they typically have about 500-600 experiments running concurrently to test new products.

One feature that encourages fairness is the push notification that anyone can sign up for that alerts you for when a new job becomes available. That notification is not dependent on your network, and they’ve noticed that this feature especially benefits people who don't have a strong social capital. They share examples like this to help product managers build more socially responsible features.

The company has also found that when networks become more diverse, it increases mobility in the labor market. For example, LinkedIn’s Plus One Pledge that encourages people to accept an invitation from somebody who they traditionally would not have connected with has been very successful in opening opportunities for those with less social capital.

On the topic of women in the workplace, she encourages women to advocate for themselves. “When I first joined Microsoft, I was the first statistician on the team. People didn't know what to do with me so I just defined what a statistician should be doing,” she says. “It’s up to you to define what kind of role you play in an organization. The more reactive you are, the more that people are going to give you orders. The more proactive you become, the better it is for the company.”

Ya has been an individual contributor for most of her career focused on solving a specific problem, but once she became a manager, her perspective shifted. “What excites me now is actually being able to help my team to be more successful. It's almost like there's no way I could solve the problem better than the way that my team can solve the problem. It really excites me when I see their achievement.”

She believes that women lead differently. “I have seen women leaders in extremely prominent positions be so humble,” she says. “I think women can be a lot more vulnerable, and it's actually a strength. When we are vulnerable in front of our team, then they relate to us. There's something different about women, and in a very good way.”

RELATED LINKS

Connect with Ya Xu on  LinkedIn
Read more about LinkedIn Engineering
Connect with Margot Gerritsen on Twitter (@margootjeg) and LinkedIn
Find out more about Margot on her Stanford Profile
Find out more about Margot on her personal website

With a prolific career spanning academia and industry, Susan’s research focuses on the economics of digitization, marketplace design, and the intersection of econometrics and machine learning. She received her PhD at Stanford Graduate School of Business, and taught at MIT and Harvard before returning to Stanford. She was consulting chief economist for Microsoft for six years and the first woman to receive the John Bates Clark medal for her contribution to economic thought and knowledge. Susan sits on the boards of Expedia, Lending Club, Rover, Turo, and Ripple, as well as the nonprofit Innovations for Poverty Action.

Throughout her career, she has built upon an early interest in auctions that she developed as an undergraduate at Duke, where she triple majored in computer science, economics and mathematics. Susan first applied her expertise to develop a market-based system for timber auctions in British Columbia that enabled a more efficient allocation of resources that was not subject to trade disputes. The system she developed in the early 2000s is still used today to price almost all of the timber in British Columbia.

While at Harvard, she was working on auction models for search advertising when she got a call from Microsoft. Steve Ballmer asked if she could come help them develop their new search engine. She had accomplished many of her academic goals: earning tenure at MIT, teaching at Stanford and Harvard and receiving this Clark medal. “I realized that this could be a good moment in my life to take a risk,” she said.

“Being a part of the birth of a search engine, and particularly the search-advertising platform was I think just a transformative experience for my life,” she said. While the Bing search engine was ultimately not able to compete with Google, Microsoft’s investment in the research yielded expertise in machine learning and cloud computing, which is now the company’s most important business.

Susan consulted with Microsoft for six years but knew she wanted to continue to pursue her career in academia. After stints at Harvard and MIT, she decided to return to Stanford as she saw it was the best place to collaborate with industry to do cutting-edge research.

As Associate Director at the Stanford Institute for Human-Centered Artificial Intelligence (HAI), she brings a social science perspective to AI questions. She explains how in online advertising, you need to understand the system and the economic incentives of the people operating in the system. An engineering perspective sees a database full of advertiser bids that feels static. An economist’s perspective sees those bids as strategic. If you understand the behavior of those firms, and their objectives, you can predict their responses to a change in the system. “It's so important to bring in multiple perspectives. There have been many cases where people have made big mistakes because they only look at it from one particular perspective.”

The effectiveness of an application is not determined by the details of the algorithm, what’s really important is that you’re optimizing the right long-term objective. The success of data science, machine learning and artificial intelligence in applications is critically dependent on having domain experts and social scientists that think about long-term objectives and how to measure them.

Currently, Susan believes there are opportunities to use technology to tackle inequality problems. She sees potential in using mobile devices throughout the world for education, training, and nudges to guide decision-making. She started the Shared Prosperity and Innovation Initiative at the business school to help social impact firms integrate more AI into their products and services.

Another technology that she thinks can help address inequality is Bitcoin. “I think I was the first economist to take Bitcoin seriously. It's fascinating from a variety of angles.” She learned how many people in the world are disadvantaged by an archaic financial system that is operated for the benefit of large businesses and banks in large countries. “If we can move money the way that we can move information, we could actually make a lot of people's lives better off,” she says.

She was used to being the only woman in computer science and economics but a lack of role models made it difficult for her to visualize herself succeeding. She felt the need to overachieve to compensate. “It's very stressful to overachieve,” she said. “But I think it translated into more accomplishments because I just didn't think that I had any wiggle room.”

Gender was not as much of an issue for her in business because she came in as a defined expert and was not threatening anybody's job. However, in academia she says the power balance is unclear, and there are no rules about who gets to choose. “Being a powerful woman actually is hard. People seem to like their women a little less threatening than I am,” she says. “When I advise women, I suggest having a clear expertise where everybody understands why you're the one who's talking.”

RELATED LINKS

Connect with Susan Athey on Twitter (@Susan_Athey) and LinkedIn
Find out more about Susan on her Stanford GSB Profile
Read more about Stanford Graduate School of Business
Connect with Margot Gerritsen on Twitter (@margootjeg) and LinkedIn
Find out more about Margot on her Stanford Profile
Find out more about Margot on her personal website

Montse Medina was pursuing her PhD at Stanford’s Institute for Computational and Mathematical Engineering (ICME) when she realized she had a great idea for a company. She left her graduate program to found Jetlore, a prediction platform that empowers retailers with AI-driven content, which was acquired by PayPal in 2018. Montse has since moved back to her native Spain as a partner for Deloitte where she is responsible for their advanced analytics and asset-enabled business.

Montse discusses her lessons learned growing Jetlore with Stanford’s Margot Gerritsen, Stanford professor and host of the Women in Data Science podcast.

She started the business with her thesis advisor and was supported by Stanford’s incubator program StartX that helps students start companies and introduces them to investors. Though they were able to raise money, there were plenty of challenges getting the business off the ground. At first, they didn't want to tell anybody what they were doing. “That already shows that we were very naïve. If we had told people, they would have given us advice, we could have adapted faster,” she says.

The technology was great but it was difficult to find the right market fit. They had to listen and pivot a lot. She had to learn how to fail, something she was not used to doing. Failure was not in her vocabulary and that was a big lesson. They had to fail to adapt, and the most important thing was to adapt fast.

She believes that entrepreneurship is also about luck. “You could be doing everything right and working super hard and being the best entrepreneur, but if you don't get some luck, you won't get the reward,” she says. And if you don’t get that luck at first, she says talk to others, listen to the environment to understand why, and then adapt fast.

Montse says it’s no longer necessary to be in Silicon Valley to start a successful company. She believes Silicon Valley is not a place; it’s really a network. For those outside of Silicon Valley who want to be part of that network, she recommends first finding successful entrepreneurs in your region as they likely will have some connections to Silicon Valley. She thinks there are going to be a lot more startups coming from all over the world.

It’s important to keep all your options open. When asked about her biggest dreams for herself ten years from now, she replied: “I don't like to box myself into a dream because that forces me into that. I think that there are so many great things that can happen to you just by letting things flow. If I were to think about a goal, then that would probably close other doors that are opening on my sides and windows that I probably would neglect.”

RELATED LINKS

Connect with Montse Medina on Twitter (@montsechka) and LinkedIn
Read more about Deloitte Espana and Jetlore
Connect with Margot Gerritsen on Twitter (@margootjeg) and LinkedIn
Find out more about Margot on her Stanford Profile
Find out more about Margot on her personal website

It was in 2015 when Margot Gerritsen was asked to speak at a data conference with not a single other woman on the program that she knew that something had to be done to get women into the field.

As she was then Director of the Institute for Computational and Mathematical Engineering (ICME), Gerritsen knew more than a thing or two about data science and became determined to change the male-dominated culture.

This determination led to the creation of the wildly popular “Women in Data Science Conference.” In putting the first agenda together, she was insistent that the conference be not about the problematic state of women in the field, but on the exceptional science of the attendees.

Now into its fifth iteration, with more than 100,000 participants worldwide, online and at satellite events spreading into six continents, Gerritsen and her co-directors of the conference have inspired women across the planet to enter the sciences and provided a platform for them to highlight their work. In addition to the conference, WiDS now includes a datathon, a podcast that Gerritsen hosts, and ongoing education programs. The results have been, quite literally, life-changing for many.

RELATED LINKS
Connect with Margot Gerritsen on Twitter (@margootjeg) and LinkedIn
Find out more about Margot on her Stanford Profile
Connect with Russ Altman on Twitter (@Rbaltman) and LinkedIn
Find out more about Russ on his Stanford Profile
Read more about Stanford ICME and Stanford School of Engineering

This podcast episode was originally published on Stanford Engineering's Future of Everything podcast.

Timnit recently completed her postdoc in the Fairness, Accountability, Transparency, and Ethics (FATE) group at Microsoft Research, New York. Prior to that, she was a PhD student at the Stanford Artificial Intelligence Lab, studying computer vision under Fei-Fei Li. She also co-founded Black in AI, an organization that works to increase diversity in the field and to reduce the negative impact of racial bias in training data used for machine learning models.

She was born and raised in Ethiopia. As an ethnic Eritrean, she was forced to flee Ethiopia at age 15 because of the war between Eritrea and Ethiopia. She eventually got political asylum in the United States. “This is all very related to the things I care about now because I can see how division works,” she explains during a conversation with Margot Gerritsen, Stanford professor and host of the Women in Data Science podcast. “Things that may seem little, like visas, really change people's lives.”

Last year, she said that half of the Black in AI speakers could not go to NeurIPS because of different visa issues. “And in that 20 seconds, that visa denial, it feels like the whole world is ending for you because you have an opportunity that's missed… Not being able to attend these conferences is much more important than people know.”

She has learned through her work with Black in AI that the number one thing we need to do is empower people from marginalized communities, which is why diversity, inclusion and ethics are not at all separate. It’s essential to have a wider group of people in the world determining where AI technology goes and what research questions we pursue. She says the industry has been pretty receptive to her proposals around norms, process and transparency because they are easier to operationalize. However, there are other things like racism and sexism where we need a fundamental shift in culture.

She has seen the potential for unintended consequences with AI research. Her PhD thesis at Stanford utilized Google maps data to predict income, race, education level, and voting patterns at the zip code level. She saw some follow up research using a similar methodology to determine what kind of insurance people should have. “And that is very scary to me. I don't think we should veer off in that direction using Google Street View.” She says she wishes you could attach an addendum to your earlier research where you talk about your learnings and your intentions for how the work be used. Timnit is currently working on large-scale analysis using computer vision to analyze society with lots of publicly available images. She says it’s critical that she also spend a lot of time thinking about the consequences of this research.

RELATED LINKS

Connect with Timnit Gebru on Twitter (@TimnitGebru) and LinkedIn
Read more about Google AI and Black in AI
Connect with Margot Gerritsen on Twitter (@margootjeg) and LinkedIn
Find out more about Margot on her Stanford Profile
Find out more about Margot on her personal website

Our WiDS Ambassadors in Paris and Beirut discuss the impact of the growing WiDS presence and communities in their countries. Christiane Kamdem, a native of Cameroon and WiDS Ambassador in Paris, is a senior data scientist at the French energy company Total where she analyzes data to create new services and improve market impact. WiDS Beirut Ambassador Lama Moussawi is an Associate Professor at the Olayan School of Business at the American University of Beirut (AUB) where she conducts research and teaches management science. 

Both women became WiDS ambassadors because they believe that role models, education, and community can make a real impact. “I believe in the vision of WiDS, which is to inspire, educate, and get educated in the field of data science, and to encourage and support more women and girls to join the field,” Lama says during a conversation with Margot Gerritsen, Stanford professor and host of the Women in Data Science podcast.

Lama grew up in in Lebanon at a time of war. “It wasn't expected that girls would go to university. Girls were expected to get married and be mothers. So when I applied to universities and I got accepted at AUB with a scholarship, it was a great opportunity that I went for and then things started from there,” she says.

She explains that young girls in the Middle East are very smart and have the quantitative skills, but they need support. WiDS provides support, guidance, and mentorship. “By showcasing role models, stellar female experts in the field, we are encouraging those young girls to not to be afraid to join,” she says. “Events like WiDS help us defy those barriers and those challenges that exist for women.” 

Christiane says the majority of the students in her grad school in Cameroon were men, and even now, she is the only woman on a team of five data scientists. After getting her Master’s degree, she started to participate in events to attract more young women in STEM fields. “It's very important to inspire the next generation, and it's important to build a kind of network of data scientists that can be models for the next generation. Because when you have a model, you want to be like this model,” says Christiane.

Both women have helped to host regional WiDS events that are making an impact in their local communities. The most recent WiDS event in Paris had nearly 250 participants. Christiane says she not only gained a lot of technical knowledge about data science, she also heard the stories from many women who had to struggle in order to be where they are now. “It was very instructive to hear that various paths can lead to great achievements,” she says.

Over the past two years the WiDS events at AUB in Beirut have gotten bigger and generated more awareness. In 2019, many more people were interested and wanted to attend the event. “We are seeing a lot of institutional support, and huge support of the local community,” says Lama. “More and more companies are participating, sending their employees, and contacting us to work on initiatives related to supporting women to join the data science field.”

RELATED LINKS

Connect with Christiane Kamdem on LinkedIn
Connect with Lama Moussawi on Twitter (@lama_moussawi) and LinkedIn
Read more about WiDS Paris and WiDS Beirut
Read more about Total and the American University at Beirut
Connect with Margot Gerritsen on Twitter (@margootjeg) and LinkedIn
Find out more about Margot on her Stanford Profile
Find out more about Margot on her personal website

Sherrie brings an interdisciplinary and entrepreneurial perspective to her research that she has developed through her work in the fields of computational finance, biomedical engineering, and computer vision. 

Sherrie explains that about 1 in 9 people do not have access to adequate food. She is using satellite imagery and machine learning to identify and map crops around the world, see where people are most vulnerable, and what interventions or policies have the greatest effect. 

“There are a lot of problems that technology alone can't solve and we still need to understand the roots of a lot of the problems. That involves talking to people in the earth sciences and agriculture and learning from them. In those conversations, data science just becomes a tool, a very useful tool but it's a tool in the context of some much larger problem,” she explained to Stanford’s Margot Gerritsen, Stanford professor and host of the Women in Data Science podcast.

Sherrie has brought her expertise to the WiDS Datathon committee where she helped develop the 2019 challenge. Participants were asked to use satellite data to identify oil palm plantations. She felt it was important to bring awareness to this issue as these plantations are contributing to the destruction of the rain forests.

Sherrie’s curiosity has taken her down many different paths so far. If there’s a field she wants to learn more about, she has never hesitated jumping in. She admits these transitions can be challenging and each one comes with a steep learning curve. While she felt overwhelmed early on in grad school, she has learned to succeed through patience and consistent effort. 

Her entrepreneurial experience running a tutoring service as an undergrad has also prepared her well for her grad school research. She says that grad school and entrepreneurship are similar in many ways.

“You're creating something new that hasn't existed before, so there are a lot of different directions you can always be going. Realizing that there are all these options out there and it's okay to be brave and pick one and go with it and see how it goes and it's okay to fail, that was a big lesson from that experience,” she says. She is excited about figuring out the new science she can explore that will have an important impact on how we see the world.

She has found her research and teaching so satisfying that she wants to pursue a career in academia. She believes that machine learning and environmental science are critical to our future. “I am optimistic that if we're honest about the sorts of problems we face, then we can collectively, creatively solve these problems. It goes back again to being patient and being consistent,” she says. “The only choice we really have is to keep trying to solve these problems one day at a time.“

RELATED LINKS
Connect with Sherrie Wang on Twitter (@sherwang) and LinkedIn
Find out more about Sherrie on her Stanford Profile
Read more about Stanford ICME
Connect with Margot Gerritsen on Twitter (@margootjeg) and LinkedIn
Find out more about Margot on her Stanford Profile
Find out more about Margot on her personal website

Ghassemi explains how she is tackling two issues: eradicating bias in healthcare data and models, and understanding what it means to be healthy across different populations during her conversation with Women in Data Science Co-Director Karen Matthys on the Women in Data Science podcast.

She says that there are built-in biases in data, access to care, treatments, and outcomes. If we train models on data that is biased, it will operationalize those biases. Her goal is to recognize and eliminate those biases in the data and the models. For example, research shows that end-of-life care for minorities is significantly more aggressive. “This mistrust between patient and provider, which we can capture and model algorithmically, is predictive of who gets this aggressive end-of-life care.”

Ghassemi is also interested in the fundamental question of what it means to be healthy, and whether that rule generalizes. It requires a different mode for data collection and analysis. She explains that the typical process is that data is generated when you go to the doctor because you are sick. However, what matters more than your infrequent doctor check-in is how you're experiencing things day to day, the self-report. She sees a huge opportunity in combining doctor visit data, self-reported data and data from wearable devices that's passively collected from people that consent to their behavioral data being used. We can use all of those different kinds of data modalities to understand what it means to be healthy for all kinds of people.

She also offers valuable insights from her career in data science as a woman, a minority and a mother. She is a visible minority because she chooses to wear a headscarf. “I became comfortable very early on with defending choices that I had made about my life. And that for me really was instrumental in the academic process. Because what is academia if not constant rejection?”

Ghassemi made the decision to become a mother while pursuing her PhD. “As a society we should recognize that having kids is not a career hit.” She felt she was able to have kids and be successful as a graduate student because there was a community around her that was supportive and recognized that having children would enrich her life and experience. She credits having a supportive mentor as being instrumental in making it all work, saying, “You have to choose the race that you can be successful at.”

She wants young women entering the field to know there is no one defined path. She says don't worry about checking boxes. Choose things that you are very passionate about. Find a mentor who's willing to invest in you, and the path you want to take. Surround yourself with good people. It's not the project that makes you successful; it's the people. If you can't trust the people around you, and learn how to work together, you are going to fail. Having the right mentors and having the right people around you should always be your guiding star.

RELATED LINKS

In addition to her job at Intuit, Lador is a WiDS ambassador in Israel, has her own podcast about data science, and is a co-founder of PyData Tel Aviv meetups.

Lador’s team at Intuit focuses on machine learning in security and fraud applications to protect customers’ sensitive financial data from fraudsters and hackers. She and her team use anomaly detection and semi-supervised methods to secure Intuit products and data. “In general, putting AI into products is not an easy task.” But she thinks we need to put a lot of effort into securing our data especially with recent data leaks from Equifax and Facebook. “I think the world is going into that direction with the GDPR and other initiatives. AI has a lot of potential of helping in that domain,” she explained during a conversation with Stanford’s Margot Gerritsen, Stanford professor and host of the Women in Data Science podcast.

Israel has a lot of expertise in the security domain because many young people study security and encryption during Israel’s mandatory military service. She had the option to do this during her service, but since she already knew she would pursue a career in this area, instead she chose to become a pilot instructor in the flight simulator. “It was a very unique experience that I would probably never get to do.”

When Lador was starting her career in data science, she did not know many people in the field. She decided to start a PyData branch in Israel because she wanted to build a professional data science community. “My main motivation was that I wanted to learn and that I wanted to have friends and people to consult with and learn from. And now I have so many data scientist friends because of all this work and it's great. I love it.”

She noticed when organizing PyData events that it was much easier to get male speakers. When she would ask a talented female scientist to talk about her work, she would say: “No, I'm not an expert… I'm not ready. I need to learn more… I was like, no, you're enough years in the field. Everyone can learn something from you.”

Being a WiDS ambassador was like an extension of her PyData work. “I get to decide what's in the conference and bring the best talks there.” Her experience organizing the PyData meetups helped her know how to create a valuable conference. She sees WiDS as a great opportunity to encourage more women to speak by giving them a platform, but also by bringing all the people together. “Seeing all those women on stage. This gives great inspiration to speak at other events, not just in WiDS. I think this is just an amazing initiative.”

RELATED LINKS

During her career at the National Security Agency, Capitol Hill and the White House, Natalie Evans Harris saw that while we collected troves of data, we didn't have strong frameworks and governance in place to protect people in a data driven world. “Data has been used to intrude in our lives. Things are happening based upon data that nobody communicated to the public was actually happening,” she explained during a conversation with Stanford’s Margot Gerritsen, Stanford professor and host of the Women in Data Science podcast.

Data ethics and responsible use of data are essentially about building trust. There's this gap in understanding what sharing data means. Two things have to happen if we're going to build a relationship where people allow their data to be used by a company. Individuals have to trust that what the company is doing with that data is something they're okay with. And the company has to be able to prove that they're being responsible with the use of the data. A company could have the best products out there, but if people don't trust you or understand what you're doing with the data, then they're not going to trust you to use the data. And then innovation stops.

She believes the biggest problem is we do not have a shared vision of what ethical practices mean. We don’t want to put broad impact laws in place to govern responsible use of data when we're still trying to define that vision. To change business practices, we have to change company expectations so that they're not only incentivized to be ethical and responsible in their business models, but they're also penalized when they violate.

Harris has been advocating for a data science “code of ethics” to create a shared vision to guide our behaviors, and then start to develop best practices around. Companies are now taking this code of ethics and personalizing it to their businesses around principles like informed consent, transparency, fairness and diversity. Companies then publicize the practices that they're putting in place to align with those principles. That's how you start to create that shared vision.

She sees that there's this transformation happening with the relationship between technology and people. For so long, technology has been this very passive thing in our lives, and now with AI and machine learning and all of these uses of data and technology, there's this tension around what technology can do and what humans should do. Until people know and understand what is happening with their data, and until companies can thoughtfully express what they're doing with the data in a very transparent fashion, we will continue to have this tension. She is hoping that this code of ethics can start to ease that tension.

RELATED LINKS

Listen and Subscribe to the WiDS Podcast on Apple Podcasts, Google Podcasts, Spotify, Stitcher