I’m interested in learning to work with some of the more popular statistical software packages but not sure where to start. From what I hear, SAS MINITAB and SPSS are the most widely used… but which which would catch the eye of a potential recruiter for ER or PE positions? Is there one that is easier, or would give me a better intro into the subject?
Matlab - register at a community college & you can probably buy a student edition for $100 or so.
I learned Matlab and it is very good…however I think SAS is more popular
Minitab is a teaching product. SAS is the real McCoy. There is no software product in the world that you can write on a resume that “would catch the eye of a potential recruiter for ER or PE positions”. If you want to be valuable in a role using statistical software, you have something on your resume that says you know something about statistics and you are able to use any statistical software. I’m sure at various points in my career I have used all statistical software. You just pick it up and staart typing…
There’s not a whole lot in Minitab that you can’t do with the Excel stat add-in. Pretty easy to use though.
Don’t put minitab on your resume unless you also italicize and bold the junior in “Looking for a JUNIOR position in investment management”
SAS is definitely the real deal. For the things it deosn’t doo quite as well (time series and higher-powered econometrics), Stata seems to be the software of choice (for academics, at least).
i think SPSS and SAS are the most widely used. others that are useful to know would be STATA, Matlab, R, Minitab.
SAS is widely used because it has nice data preprocessing features that are useful with the often *very* large datasets used in quantitative applications. SAS tends to be the most popular as far as I’ve seen; the main problem is that the license is fairly expensive, especially for small shops or individuals. Even student software tends to be highly restricted with quickly expiring licenses. I used SPSS in teaching because that’s what I learned in graduate school. It’s relatively simple to use and has the advantage that it doesn’t take long to figure out the syntax. However, I don’t find it terribly well suited for financial applications. You can certainly use it, but its bells and whistles are designed more for sociology, psychology, and political science applications, where it is good. Stata is fairly powerful and relatively inexpensive compared to SAS. Stata’s business model appears to be to sell the software relatively cheaply and charge a ton for documentation. However, you can slog through it. I tend use Stata for statistical work with financial datasets. Matlab is very nice when you need to do matrix operations or create efficient frontiers. Supposedly, you can do statistical work with it too, but I tend to do lots of data cleaning and transformations and find that it doesn’t handle data labels very well, which makes it a pain in the butt to figure out what variables you’re using in your models. So I Matlab very much for it’s abilities to do matrix stuff, but wish it’s data management features were more sophisticated or at least user friendly. BTW, Stata has a matrix manipulation language called MATA which I haven’t tried. R is similar to Matlab and has the advantage that it is free. (Search for “R” on Google and you’ll find it). The main problem is that the documentation is frustratingly sparse. However, if you’ve used Matlab and don’t feel like buying a license for your own use, R is not a bad substitute. Mathematica is good for symbolic math. I used it a few times and was very impressed; however, I haven’t done enough abstract stuff to really make use of it. I’m sure people who do heavy derivatives stuff find it great. Every time I use one of these programs, though, I find I have to relearn the syntax, which is a bit embarrassing, since supposedly I’ve already used these things.
Forecast Pro is pretty good for ARIMA, Regression/Dynamic Regression, Expo Smoothing, and other popular statistical forecasting tools. It’s pretty easy to use too. I know Boeing and a lot of other large firms with long planning horizons use Forecast Pro in their repertoire.
STATA, Eviews, Matlab
Thanks for all the feedback, I’ll definitely start looking into SAS. I’ve already been exposed to Matlab in my engineering undergrad but mostly in modeling systems/circuits; any suggestions for a good book on Matlab that has a statistics focus?