Diversity in data science campaign launched

Toni Sekinah, Research Analyst & Features Editor, DataIQ

Kim Nilsson, CEO and co-founder of data science hub Pivigo and its associated training provider S2DS, was attending a lot of events that discussed the importance of diversity in her industry when she had a revelation. Fifty-four nationalities are represented among the S2DS students and 35% are women.

Nilsson, in an interview with DataIQ, said: “Our programme was already quite diverse and we wanted to showcase that.” Diversity in the workforce is an issue that has been brought to the fore as the benefits of it have been widely documented.

To do that, a new website was launched at the beginning of May, www.imadatascientist.com, which hosts interviews with three graduates of the S2DS course. They are encouraging data scientists from around the world to do the same and submit their personal stories of how they got into the industry. “It’s really about engagement. We want everyone who feels that they are a data scientist to tweet #imadatascientist and tell the world they are proud to say they are one,” said Nilsson.

Kim Nilsson, CEO and co-founder of data science hub Pivigo

Nilsson and her team want to get the message across that you don’t have to be a 25-year-old, white, British male to become a data scientist. “We think this could be a really good initiative to build interest in the non-traditional communities for data science careers,” she said.

The campaign has already been successful in gaining high-level support from big data influencers, including Kirk Borne and Renee Teate. Nilsson also got in touch with DJ Patil, former chief data scientist of the United States Office of Science and Technology Policy, who suggested that she get more organisations supporting it so that it becomes a cross-industry initiative.

“We hadn’t thought of that, but it sounded like a great idea.” Nilsson and the team are now contacting training organisations, as well as big service providers like Horton Works, Cloudera and Microsoft, trying to get them on board. “The goal for us now is to get as many organisations involved as possible and really make it a global, cross-industry initiative,” she said.

But why is diversity important in data science? According to Nilsson, the main reason is the risk of bias in algorithms. She cited a case in the US where software used to determine how likely it was for a defendant to reoffend was found to be biased against black people. Black defendants were wrongly labelled at twice the rate of white defendants as likely to commit another offence in the future.

She said: “Those sorts of biases can be really hard to spot, especially if the people working on the algorithm are blind to the biases themselves. If no-one questions the data that goes into it or the output that comes out of it, there’s a grave risk of biases being introduced.”

Another reason is innovation. “You have to have people who can come up with new and different ideas and who are not all thinking the same. That is just as relevant, if not even more, given that data science is a such an innovative industry,” she said.

Nilsson argued that her industry has already been making strides in the right direction in terms of gender diversity and estimates that the proportion of female data scientists has increased from approximately 12% to 25%. This could be the result of greater awareness of data science as a career among the academic community, something else Nilsson has noticed.

The groups that Nilsson hopes will be encouraged to take up a career as a data scientist include disabled people and older people.  “There is no reason why disabled people, once given the right tools, should not be able to do data science jobs,” she said, pointing out that the head data scientist of indeed.com in the US is blind.

She also said that it is never too late to change careers and S2DS has trained individuals in their 40s and 50s wanting to make a career transition. Highlighting the benefits of a career in data science, Nilsson said that it is a fun and challenging job, especially for those with an academic background, and one that has a supportive community around it. It also has the benefit of having a large job market which means that, “if you become a data scientist, you are basically guaranteed a job,” according to Nilsson.

She is certain that the campaign will run for a long time as she knows that it is not going to solve all of the problems overnight. However, she is positive that, given time, things will change. Nilsson hopes she will see some of that change reflected in the type of applicants she will see for future S2DS programme.

“Ultimately, that is the goal - to have a more diverse workforce. Anyone can be a data scientist if they want to be. I think we need to be very positive about this industry. I definitely am and it is a great future for anyone that is interested in going into it,” she said.