The Case for More Women in Data Science

Editor’s Note: From time to time we like to feature guest bloggers who we feel can bring an interesting perspective to our readers. With the “Grace Hopper Celebration of Women in Computing”event kicking off this week, we wanted to share an exclusive post written by Claudia Perlich, Chief Scientist at Dstillery, who we worked with closely when we partnered with the Knowledge, Discovery & Data (KDD) Mining conference in August 2014. Claudia currently serves as the first female chair of KDD and we thought her story was appropriate to share as we continue to celebrate, support and encourage talented women to push the limits of computing every day.

By Claudia Perlich, Chief Scientist, Dstillery

I have to admit that I never really gave the number of women in data science much thought. Maybe it was because, by some lucky accident, all three of my NYU faculty advisor’s PhD students happened to be female. And about half of my predictive modeling group peers at IBM Research were female. And half of the PhDs here at Dstillery are female. Or maybe in the bigger picture, having spent the better of 15 years in computer science departments, research labs and weight rooms, being around mostly men was perfectly normal – in fact, expected.

The number of women in data science didn’t get my attention until about one year ago when I was asked to be the general co-chair of one of the biggest and most well-established DS conferences: SIGKDD 2014 (KDD).

I was thrust into the role of having to argue that “we need to get more women speakers.” There are many excellent women in my field (I know many of them personally), but for the most part they are not on the radar for keynote speeches and rather few of them have titles like Chief Scientist. Why do they seem rather invisible?

One response is the personal choices and trade-offs these women make. Most of my female friends have chosen to stay in academia – some because they enjoy teaching, some because it seemed like the thing they were expected to do, others because there used to be really good immigration reasons to choose academia, and others because it promised a predictable future (twelve hours of work a day for six years, and then hopefully tenure without having to move around and look for good school districts). Ten years ago, having an advanced degree in the equivalent of data science was not exactly sought after in industry, and few of us ventured in that direction.

The problem with data science in academia is that that’s not where the magic happens. It happens at the Googles, Facebooks, Microsofts, Bloombergs and IBMs, as well as startups like Dstillery. This is where the richest data and most interesting problems are – and it’s not accessible to most academics. In fact, for them, getting access requires heavy networking.

In this regard, I got lucky. I was turned down for a job at a good business school and wound up at IBM Research, the only non-academic job to which I applied. I found an environment that was immediately appealing: collegial, open, not at all political. The department was run by one of the most impressive women I’ve ever known, and she sent me a personal note to congratulate me on the birth of my son before I had even accepted the job. I felt at home. I took the job, and never looked back at academia.

After my 5+year stint at IBM, I became Chief Data Scientist of Dstillery – formerly Media6Degrees. It was an opportunity I couldn’t pass up.

One clear lesson from my journey is that having women already in the environment, preferably in leading positions, can lead to a healthy culture that will encourage other women to join.

I hesitate to make sweeping generalizations, but many attributes that I equate to my gender have helped me in my career as a data scientist including communication, intuition, flexibility and pragmatism.

While I can’t claim that other women in my field share my perspective, I can say that many share my passion for increasing our numbers. It’s risky to list specific names as you inevitably leave deserving people out; however, women such as my colleague at Dstillery, Lauren Moores, Brenda Dietrich, IBM Fellow and Danah Boyd, principal researcher at Microsoft and founder of the Data and Society Research Institute, are very active in encouraging and mentoring young females to both “speak” and think data.

With the percentage of female registrants at SIGKDD (or KDD, the most well-established global data science conference) only reaching 23% this year, companies like Microsoft are taking a leading role by sponsoring key events to drive higher participation such as the annual  KDD Women of Machine Learning Breakfast.

Likewise, Bosch helps foster mentorship, guidance, and connections of minority and underrepresented groups in data mining, while also enriching technical aptitude and exposure. The company actively invited minorities to KDD by providing travel and registration support for a workshop titled, “Broadening Participation in Data Mining.”

And, Bloomberg L.P. continues to believe that data is crucial to making informed decisions that can have a broader societal impact. The company hosted a portion of the KDD conference at its main headquarters in New York.

These are very encouraging signs. The ultimate scenario, of course, would be gender parity or, more preferable to me, true gender blindness. Our goal is a world where women leaders come easily to the mind of data scientists of both genders. Obviously, we have a long way to go. But if those of us who have blazed the trail for women in data science continue to recruit and guide new members to our ranks, we can get there.