It's been about two years since I started taking the Coursera data science specialization courses. I completed I think six of those courses. I still have the Regression Models, Practical Machine Learning and Developing Data Products courses to complete before doing a Capstone project. I'm not sure if I can still qualify for the certificate given how much time I've taken.
In the past year I have been so busy with the development tasks I have at work that I have not been been doing any data science projects. I'm seeing others get into it, however, and it is making me re-think my near and longer-term future.
With retirement planning also on the board, I'm thinking that it would be nice to not have to retire completely -- that would be pretty strange, going from full-time work to no work at all in six years or so. Even if we completed the plan that we are working out with the financial consultant, it is dependent on my earning my current salary and benefits until retirement age. I'm just not sure I want to spend the next six years doing software development the way I have in my career.
Doing data science project-oriented work for various organizations appeals more to me. It would involve:
- client interfacing
- technical thinking and doing
- involvement in interesting and varied subject domains
- writing, whether in the form of reports, articles, books and blogs.
To pick up the data science discipline again, I think it would be great to work with the book club group we are starting at work, beginning with Introduction to Statistical Learning. That's a challenging but well-written book from which I've already read two or three chapters.
I've also signed up for Renee Teate's Data Science Learning Club which she combines with her Becoming a Data Scientist podcast. Renee frequently tweets interesting data science articles written by others and is a tireless student and practitioner of data science.
One big decision I need to make is whether to continue with R or to brush up on Python and learn the Pandas and SciKitLearn libraries. R has memory limitations which could impact projects I might work on. On the other hand, the book club group is probably going to be using R. That doesn't mean we couldn't individually use different languages since the language implementation details are not as important as the concepts and algorithm techniques.
There are plenty of Python developers in the data science field. Renee, mentioned earlier, as well as some of the people organizing and attending Portland area data science meet-ups.
I think I'm going to give re-learning Python a try, just to have that in my toolbox. I like the idea of having a general-purpose programming language to use for data science projects.