The 3rd incarnation of the Data Science Summit here in Redwood City just wrapped up. It was an impressive group of speakers representing a diverse set of industries: healthcare to MOOCs to wearables. Serious data science has clearly permeated many realms where data existed but never got the attention it deserved.
However, when it came to gender and minorities, we got a seriously homogeneous view of practioners. Out of 49 distinct presenters/speakers there was a sum total of 4 women panelists (Annika Jimenez, Christina Farr, Pek Lum, and Monica Rogati) who presented onstage, and zero women keynotes. The audience was no more than about 20% women. I counted zero African American presenters. These stats are not at all surprising, but they are still a bit shocking.
A roundtable discussion
At the Summit, I ran a special lunch discussion with the following question: “How can we increase the pathways and opportunities for women and underrepresented groups in data science careers?” There were 10 people who showed up for the discussion, 2 men, 8 women. The desire for role models (in company leadership positions, as mentors, etc.) was a critical point for all the women as they considered their career paths.
Usually we think about challenges that unrepresented groups face as they progress in their careers, exposing too many off-ramps that derail and dissuade. But the most enlightening part of discussion for me was in hearing about the many paths people took to get them into the data science world. Some started in traditional STEM fields and then caught the business bug. But many started in non-STEM fields, like communications and political science, and found interest in computational science and machine learning later on in life. They took online courses in statistics and CS to build the techinical aspects of their DS chops.
I never really thought of data science as an attractor with multiple viable paths. Of course all of this depends on what precisely our definetion of data science entails. Women in the lunch session talked about how they saw data scientist capability partial as the “art of story telling about data”. Many will also see need for some particular skill set in statistics, computer science and/or machine learning.
What can we do?
The most obvious thing we can do to increase participation is the dedicated education, mentorship, and promotion of those that are not traditionally represented proportionately in our fields. To that end, today, wise.io, the company I co-founded, sponsored the Women in Machine Learning Workshop which took place in Lake Tahoe, Nevada. An amazing member of our data science team, Erin LeDell, was there to present and represent.
We can and should take deliberate steps. It’s not just the right thing to do, it’s also self-serving: success in data science activities comes in large part when we bring diverse views and life experiences to the table.