Two very good posts:
- Jamie Hall, 3 data science skills that economists can use
- Chad Orzel, Planning To Study Science In College? Here’s Some Advice
As stated in the title, Orzel’s advice is intended for college students. His recommendations:
- Learn to do algebra
- Learn some statistics
- Learn to program
- Learn to communicate
- Get involved in research
In contrast, Hall’s advice is geared towards people who’ve been working in the field for a while. His priority recommendations:
- Source control for everything, all the time
- Cross-validation
- R and Python
I am learning (the hard way) the value of source control. I’ve been incorporating cross-validation in my work for a while now and I am continually learning new things about it. Hall’s recommendation to learn R and/or Python make sense. Based on my conversations with undergraduates and recent grads, Python is becoming the lingua franca of technical computing. (R looks great for statistical analysis. I don’t know enough about to gauge its utility for general scientific computing.) My sense is that MATLAB has been the lingua franca for a while now but that it is being displaced. (That Python is free and that MATLAB licenses are costly no doubt contributes to this.) I write “my sense” because FORTRAN was the standard when I was an undergrad and until recently I’ve done most of my scientific programming in IDL. That stated, over the past few months I’ve been coding in MATLAB rather than IDL because a) I need to improve my MATLAB fluency and b) most of colleagues are more fluent in MATLAB than IDL. I made a half-hearted attempt to learn R earlier this year and will probably give it another go. Finally, I am curious about the basis of Hall’s statement:
Working in Python encourages you to write clean and reusable code, in the way that EViews and Matlab encourage you to write bad code.
Maybe I’ll send him a note.