female data scientist, woman leading team

Good data science requires diverse data scientists

Article by Justine O’Neill, director, Analytic Partners

female data scientist, woman leading teamAsk someone to picture a data scientist and what do you think they are most likely to conjure in their minds?

Somewhat depressingly, I’d hazard a guess that they would imagine a man, and quite possibly a ‘geeky’ man. In some ways they wouldn’t be wrong – most data scientists are men.

According to the Boston Consulting Group’s (BCG) research carried out earlier this year, only 15% – 22% of data scientists are women. It doesn’t have to be this way, but it will take concerted effort on many people’s part to change it. About 55% of university graduates are women, but from that point onwards the funnel narrows. Only about 35% of STEM degrees are held by women and that drops in the workplace with around 25% STEM jobs being done by women.

So, how to unpick this challenge? Firstly, there is a distinct image problem for data science, algorithms and artificial intelligence (AI) in general. There has been plenty of hype around AI and how it will quickly answer so many of society’s problems and automate mundane and labour-intensive roles – but the media is also full of stories about its limitations and the problems that have arisen when too much weight is placed on algorithms at the expense of human insight.

Most recently we’ve had Ofqual’s disastrous algorithm for A level and GCSE exam results which tipped schools and universities into chaos and was met with derision from teachers and students alike before the government was forced to ditch it entirely. Oh, the irony as we try to make a case for improving data science’s image and appeal among this cohort of students.

But the other challenge centres on data science’s lack of appeal for women specifically. This seems to be partly because when assessing career options, female STEM students are looking for applied, impact-driven work – they want their jobs to have a tangible effect and don’t see data science as fulfilling that.

There is clearly a job to be done among all businesses looking to hire graduates to explain more clearly how data science solves business problems – to promote its demonstrable attributes. Everyone working in the industry should share their inspiring stories about the rigours and rewards that come from their jobs. Students want to hear specifics and get to grips with what the day-to-day expectations and experiences of this job would be – show why it’s not just the domain of the nerds.

Diversity in our sector is imperative. As the author Margaret Heffernan says, “algorithms are opinions encoded in numbers” – we need the broadest range of voices building and working on those algorithms to be alert to the bias that can be built into the data sets used to create them. If your team has genuine breadth of thought and experience, then it is more likely to identify biases and produce more accurate and balanced results.

The business case could not be clearer – ensuring a company has diverse teams is not just because gender balance is a ‘good to have’, it’s essential for strategy and success. It is why men should be championing diversity with the same enthusiasm as women.

No one says this is easy. It may require a root and branch rethink of how your organisation fills its roles. I suspect many people involved in hiring data scientists will bemoan the disproportionate number of men applying to women for every role. But even from this starting position, businesses can successfully achieve a better balance in their workforce.

Everyone needs to look at the wording of their job ads, the tone of voice used and where they are placing their adverts. If you use recruiters have you explicitly requested more diverse longlists of candidates? Look at who internally is involved in the interviewing; changes can be made at all points to help nudge toward a more balanced workforce.

I work for a global analytics consultancy where three of our senior team are women, starting with our president and CEO. This is not the case for many of our competitors, but it does show how diversity can be possible.

A shift is taking place – clients want to work with more diverse agency teams, conferences want better balanced panels and speakers, younger candidates want to be in organisations that better reflect the world outside of work. The business and moral argument are aligning, and everyone needs to get on board.


If you are a job seeker or someone looking to boost their career, then WeAreTechWomen has thousands of free career-related articles. From interview tips, CV advice to training and working from home, you can find all our career advice articles here