This obviously is something all (or most) PhD students need to know but what would be the essentials needed. I already can work with MATLAB pretty well and HTML I know quite well. I will learn STATA in the fall when I have my econometrics class this fall but what else am I missing in terms of essential/required programming skills for a PhD. Would love to hear the diverse skillsets that you guys have!
C++ is very good especially for large data set.....in MATLAB unless you vectorize your code it can be very slow sometimes even when vectorized, try to avoid for loops. I wrote a trading engine using C++, I tried to do the same using MATLAB but gave up its just too slow.....But I think in econ Phd you dont really need C++, S-plus, stata and MATLAB are good enough for econometrics and macroeconomics.
I started out as a CS major in college, though that was many years ago. I haven't done any serious programming in quite some time. However, I think my list is Basic/VB, Turbo Pascal, C++/V++, Java Script, C-Shell Script, HTML, and a bit of SQL. I'm pretty rusty, but its all in there somewhere.
Now something that I wish I had was FORTRAN, as apparently its used a bit still. Go figure.
University of Wisconsin - Madison: Took my Masters and ran.
Depends on what you're trying to do. If you're asking with an eye towards classes, I'd say don't bother. If you've learned Matlab and will have learned STATA, you'll be able to learn whatever you need on the fly. Programming isn't rocket science.
If you want to do agent-based modeling (which I think will get used more and more in the next 10 years as a methodological tool), you should learn Java. C or C++ is still the fastest if you have to do some very intense computations. FORTRAN might be quicker, I don't know. I know mathematicians love FORTRAN.
I would also be useful to know Mathematica or Maple, since they are good for symbolic mathematics (like differentiating something complicated, or dealing with combinatorics). Combinatorics might sound a bit off, but I think it's important in certain areas of mechanism design/matching theory, and also if you want to deal with networks of agents interacting.
Yea my basic interest is Monetary Economics...I figured what I know would be enough but always great to hear what other people think
You should be in pretty decent shape. MATLAB is used quite a bit. (Though as quantwanabe pointed out, C++ is usually superior if you have large data sets.) Learning STATA in an econometrics class will give you some of the basic tools, but it won't help you much when it comes to actually using STATA for manipulating data and cleaning data sets, which I think it its main strength anyway. (Data almost never come in a workable form.Originally Posted by SunnyDutt
)
UCLA's statistics department has a set of good tutorials for STATA available at: http://www.ats.ucla.edu/stat/stata/default.htm
It also has some videos that show how to do stuff. I remember always telling myself I would watch them eventually, but I never got around to it. Maybe this summer!
MIT Economics, class of 2011
This thread looks really interestingI am willing to specialise in Financial Econometrics; so do you think that Matlab and Stata would be sufficient? Or would proficiency in other softwares such as C++ or Java be necessary? Financial econometricians work with pretty large data sets, but then I don't really know whether these sets are "that" large
S-plus with Finmetrics is pretty book for financial econometrics, its library is great and a lot firms on wall street use it. but if you can invest in C++ do that, you will not regret at first it might be hard but its worth it
MATLAB and STATA are necessary for almost everyone (if you never plan to work with data at all and will only do theory, you won't use STATA much.) For applied micro people or anyone doing a lot of work that involves combining large data sets, SAS is extremely useful.
Some macro people still use R (also called GNU-S; gotta love those clever programmers). Time series people often use RATS.
You'll find some professors who still write FORTRAN code, but most grad students doing work that complicated (and younger professors) do it in C++.
If you know MATLAB and a little bit of STATA going in to your first year, you'll pick it up as you go. The best way to learn these things is to use them. Once you have one figured out, learning another becomes mostly about syntax (MATLAB is a bit different). If you were going to take a class, I'd recommend C++ simply because you'll learn logic that is extremely general and you'll also learn good programming habits that you will carry over to your work in other packages. But learning C++ before (or, probably, during) grad school is NOT necessary.
There are currently 1 users browsing this thread. (0 members and 1 guests)
What you can do
LinkBack URL
About LinkBacks
Reply With Quote
Bookmarks