Since I’ve started on my Ph.D. course work, I found that I am having to do a lot of statistical work. I figured there would be some but there’s a lot more than I expected.
My school, like many others, has standardized on SPSS for statistical work. However, being the FOSS guy I am, I chose to find an open-source stats program to use. Plus, I didn’t want to spend $200+ on a program that expires in 4 years.
I found R, a FOSS statistical package that is surprisingly similar to Python. R is based on the S language, so it has a lot of programming features available. But, like Python, most of the necessary functions are built-in; rarely do you have to do any custom work.
The hardest thing I’ve come across so far is figuring out how to output my plots the way that I want. Well, another hard thing is learning some of the more “esoteric” statistical aspects, such as ANOVA and stepwise simplification of multiple regression models.
Overall, however, R is a very capable environment and the cost is great. For less than half the price of SPSS, I was able to buy 3 books about statistics and R so all I need is time to read them. The books and online documentation are usually sufficient to answer any questions I have; for anything else, there’s always Google.