Are You Going To Be A Data Scientist? Or "Just" An Analyst?

I came across this post today which I thought was a great proxy summary of something which has been niggling at me for a while. Just how familiar are the frustrations Dr Penny Wilson lays out here in proud defense of the talents of GPs to those of analysts who seem to have become fair game for accusations of somehow being "fake data scientists?" I read the article and replaced specialist with data scientist and GP with analyst and it really did bring home how we're in danger of under valuing the bedrock of great insight/analysis teams and their skills. I've edited Dr Wilsons words in the next few paragraphs. I think it makes an interesting read.

"The ongoing debate about data science skills seems to imply that “analyst” and “data scientist” are two diametrically opposed alternatives, and that analyst is the lessor of the two. If you’re smart, ambitious, passionate and successful you become a data scientist. If you can’t get into anything else, or if you want the easy option, you become an analyst. It’s seen as a back-up option, not as a worthwhile career in itself.

The data scientist vs. “just an analyst” dichotomy also perpetuates the idea that analysts are not “experts” in their own right, or that analysts are the amateur statisticians that do the easy bits and then refer on when it gets too complicated. Analysts ARE experts.

So where do all these negative attitudes come from? Unfortunately some commentators perpetuate these views. They refer to analysts dismissively as “fake data scientists” and I have heard some loudly criticise analysts for not knowing everything about a specific field, apparently unable to appreciate the enormous breadth of knowledge the analyst has in other areas.

The reality is that analysis is an enormously rewarding, challenging and varied career and that no two days are ever the same. Analysts often have no idea what is going to walk up to their desk next and every business problem comes with added layers of complexity from stakeholders personality factors, business circumstances, political situation, expectations, prior knowledge of data/analysis etc. Not only do they have to be able to initially manage every single business issue imaginable, they have to be able to do it without a full suite of analysis tools/data and always to deadlines way short of ideal. What’s more, they are not managing that stakeholder just for that issue — they are helping them in the short, medium and long term.

Don’t get me wrong. I have a huge appreciation for data scientist/specialists. In particular I appreciate the depth of knowledge and skills that they have and I know that they, too, want the best outcomes and that the system works best when everyone works together. However, I’d really like for people to stop asking, “Are you going to be a data scientist or just an analyst?” and instead to ask, “What type of data scientist are you going to be?” It’s up to all of us to change the conversation and give analysts the respect and prestige they deserve, so that all data career choices can be seen to be of equal value."

My own view is that for every specialist we need in machine learning, AI, deep learning, NLP etc most insight/analysis teams probably need 9-10 "general practice analysts" (based on the estimated skills shortages from the original McKinsey Big Data report). Recruiting great analysts isn't helped if we talk about the role as somehow not being the equivalent of a "proper doctor."

I think Dr Wilson may have accidentally written the analysts rallying cry

"No, I’m not “just” an analyst. I’m a broadly-skilled, sub-specialised, expert data professional, providing a damn fine analysis service to my stakeholders and my business. And I absolutely love it."

https://www.huffingtonpost.com/penny-wilson/myths-about-general-practitioners_b_3937618.html