Dear HN,<p>I’m just about to finish a PhD in formal logic for multi-agent systems. I don’t want to stay in academia and I’d like to try a Data Scientist career. My background is in philosophy, and even though I taught some computer science and maths courses, my programming skills are those of a mediocre undergraduate computer science student’s (at best). I know Haskell, Scala and Python a little bit, and I’m learning R. I have never taken part in any serious programming project, though. The biggest program I ever wrote had less than 300 lines of code.<p>So here’s my question to you as a community of experts. What is the best path for me? Should I simply learn Java and do exercises from some textbook one by one, and then study machine learning, hadoop and the like? Should I rather start with SICP and a good algorithms book? Or should I perhaps take a Data Science specialization on Coursera (I’m trying it, but my coding skills are too poor even for their elementary R course)? Something else?<p>Any advice is much appreciated.
Disclosure: affiliated with the program[1]:<p>We offer both a 12 week immersive in data science (with structured curriculum augmenting a focus on practical data science) and a 6 week project focused curriculum leading up to a job fair. This is a python focused curriculum focusing on scikit learn.<p>Outside of that, I ended up becoming a data scientist and engineer via taking on consulting projects and working my way up on difficulty. There's a mix of online materials out there you've already mentioned that work more than well enough. My only other piece of advice is picking a concentration and becoming as good as it as you can.<p>1. <a href="http://www.zipfianacademy.com/" rel="nofollow">http://www.zipfianacademy.com/</a>
I would join this program. <a href="http://insightdatascience.com/" rel="nofollow">http://insightdatascience.com/</a><p>Note, I am not affiliated with the program, I just know that they do exactly what you're asking, and have produced amazing results.
I highly recommend the intro to data science course on coursera:
<a href="https://class.coursera.org/datasci-002" rel="nofollow">https://class.coursera.org/datasci-002</a><p>Even if you have only a little programming experience if you spend a fairly reasonable amount of time on this you'll come out with a lot of practical knowledge which will allow you to jump into more advanced courses. (It has a bit of python, SQL and R).
If you want some simple (maybe!) real world projects to test yourself on then get in touch using the email in my profile.<p>I do some work for a charity trying to improve things using data. I have several smallish, self contained data science projects that fit in the important but not urgent quadrant.