Hi all! I am thinking of transitioning to a data science role. I am currently a statistician in the pharma/life sciences space (think clinical research/trials), but I no longer feel this is the direction I want to take my career, due to (relatively) low compensation and limited career growth. I have a master's in statistics from a respected university. In terms of programming, my current role primarily uses R/SAS/SQL, but I have working knowledge of Python, too. In terms of statistical methods, my role focuses on experimental design/ hypothesis testing/ traditional stat methods. I also have some training and experience with ML models (though, admittedly, I do not have much experience with more advanced topics such as DL/RL methods). How feasible would it be for me to transition to a more typical data science role? I am concerned that my lack of knowledge regarding ML and general business analysis may be a hindrance. If such a transition is reasonably feasible, then: 1) Would I need to take a temporary pay cut to make it happen? 2) Are there certain types of DS roles that I should be prioritizing for applications? 3) Which topics/resources should I consider to bridge any gaps and make myself an appealing candidate for hiring managers? Thanks in advance. YOE: 3.5 TC: ~110k
I’ve found statisticians make the best DS due to their deeper grasp of the assumptions behind data models. They hit the sweet spot between the SQL monkeys and import scikitlearn bros. Should be easy for you if you market yourself right. The biggest issue I’ve found pure statisticians face in business is querying enterprise databases though that’s hardly rocket surgery. Python is currently de rigeur mostly due to the aforementioned import scikitlearn bros even though R is actually the superior data science tool. I still operate with mostly R. You’ll find statistical analysis doesn’t have to be as tight in business as in academia or research. Personal experience here but biostats/life sciences maps really well to product or marketing DS. People who know what a split plot design is can save their companies millions. I don’t think you need to step back in pay for a DS role. Learn a little bit of SQL because that’s been a part of the technical screening for any DS interview I’ve had
I wish I could upvote your comment like 420 times
Thanks for this post! This is validating to read. I was definitely planning to brush up on my SQL skills, and will think carefully how I can craft a narrative behind my experiences.
Your TC is low you can't cut it down further . Just try python focussed DS roles and you can actually make a TC jump
If it’s product analytics, then you wont need to expect anything new but if you’re looking for ML-focused engineer/researcher role, than you’ll need to Leetcode in python and demonstrate a bit in DL.
Most DS people don't have a DS degree. Just be smart and know Python and SQL. The rest takes care of itself.
Sql and python is good combinations. Just prepare well and crack a good interview and join a good company
Just do a certificate in DS you’ll officially be Data Science!!
Depends what kind of DS role. Product Analytics is less about hard skills and more about experience working with product teams
Let me put down some problem statements for you to work on to move forward in data science in the area relevant to you are currently working on:- 1. Clinical trial site and patient identification with less dropout ratio and potential targets . 2. HPLC optimization for SMDD development -- what can you do there with data science? 3. Social media ADE /ADR analytics for pharmacovigilance 4. HCP identification for drug promotion campaigns.
You are already doing data science
Came to say this. Your already doing DS but just with bio data. Just give yourself the title if that's really what you care about.