In defense of the non-Engineer Data Scientist
A case for the octopus over the unicorn.
“Data science encompasses a set of principles, problem definitions, algorithms, and processes for extracting nonobvious and useful patterns from large data sets”. Nice quote taken from here.
I was going over some MLPL (Machine Learning Process Lifecycle) materials when a familiar feeling hit me. You see, my background is weird, I have a bachelor’s and master’s in Business Administration and little later built the courage to get my Physics degree as well, it was a dream of mine that for a long time I thought I wasn’t enough of a “genius” to fulfill (that is a whole other story on impostor syndrome). The thing is, MLPL has a lot of similarities with strategic planning, which is my favorite part of business administration.
I would go as far as to say that many methodologies, like CRISP-DM, SCRUM and other agile as well as lean methodologies are in fact applications of this strategic thinking that I feel very comfortable around due to my administration background. Physics, on the other hand, gave me the chance to understand complex mathematical concepts and be ok with working on things like these. After all of that, learning about data science was as magical as seeing a full double rainbow.
My point is, there is a lot more to data science than programming and software development. We’ve all seen the famous Venn diagram for data science, and the below version which I prefer. The search for the mythical unicorn has led businesses to create jobs with almost impossible requirements, not knowing that the original Venn diagram is not necessarily describing a person, but a team effort. I am not saying that there aren’t any unicorns out there, but their development is certainly not a straight path, it takes time, experience and involvement. They started somewhere, not being very unicorn-ly.
This piece by Thomas Davenport has an interesting approach to the issue at hand, denying the existence of the unicorn among other things. He says:
“ …the data scientist takes on a variety of different forms and specializations: the statistical data scientist, the computational data scientist, the strategy consultant data scientist, and so forth... when I interviewed more than 30 data scientists almost a decade ago for an article I found that the most common academic background was experimental physics, but there were also data scientists with backgrounds in astrophysics, statistics, sociology, meteorology, artificial intelligence, and many others”.
In reality, we can’t expect everybody to have had a quarter life crisis and go back to school to study a super mathematical subject after having a very business oriented degree and on top of that learn ninja level programming (which I am trying my very best to do, imagine me going back to school now for yet another degree because I am not a software engineer …eye roll). I don’t know if I want to try to be a unicorn anymore, something tells me it’s not healthy. I propose the existence of the data science octopus! Unlike the unicorn who is perfect at everything, the octopus can do a little bit of a lot of things and tries its best.
When I look at the different entry level job postings looking for a unicorn with (only) a Computer Science degree, I can’t help but cringe a little. The realistic option would be to have a couple of areas checked and have training for the third one or at least patience while skills reach the desired level, because this would be beneficial for all parties involved. Bottom line is, the position is attractive to companies because an individual who is able to quickly understand the business, has a high level of curiosity, is math-oriented and learns how to use the right tools can bring a lot of value to any project; but these things aren’t all determined by one specific degree.
To me, data science is a place where my all over the place passions come together and thrive, where data obsessed people from all sorts of backgrounds can work together to solve problems and build solutions. There’s value in that, but a paradigm shift is needed if we are to make the most of what the field is here to offer.
Further reading on this topic.
