I have seen many jobs asking for basic knowledge of SQL/JAVA/SPSS/SAS/MATLAB.I know Basic / Advance exel+VBA.Can you guys suggest me if i really need to know above mentioned programming skills to advance further in my career or find a job in risk management??I am currently working as assosiate ER in a MNC.
From my understanding, all the positions requiring MATLAB, SAS, SPSS, SQL, etc. are quant jobs. So, if you’re interested in a career in quantitative analysis, it would be beneficial learning more advanced programming and system applications. If not, you probably don’t need to worry about it.
how about get a gf and go to town once
In India even some ER jobs require you to have SAS/Spss skills…
I will probably learn SQL and SAS with in next year.
Equity research is generally fundamental research not requiring programs like SAS/SPSS…now if you’re going to be doing equity derivatives research…then learn MATLAB, C++, VBA, and SQL. SAS and SPSS are utilized more by academics in other kinds of research so I don’t see why in India ER jobs are asking for it…seems silly.
LOL
If you are already in finance, probably MATLAB or R is most useful, since you can do data analysis with it. You are unlikely to be a major coder if you are already on track for a finance career and not already doing coding. I think you said you already know VBA, which is probably the most obvious choice for traditional fundamental-oriented research.
SQL is also useful, and not too hard to learn, but really only useful in grabbing data from databases. Most coding projects requires other stuff too, so it’s rarely a language that you’d want to know without some other language on top of it. Also, if you are just going to read data from sources, SQL is a lot easier to learn than if you are going to be reading and writing and building databases, which involves a lot more complicated thinking and analysis.
If you are still trying to break in or unsure that you will succeed, Java is probably best of those on your list, because it can be used in many other lines of work too. C++ is kinda the gold standard for coding, in that if you know it, people will assume you can learn anything necessary. C++ is kinda interesting in that it will fit many different coding paradigms (functional, procedural, object oriented, etc.), but it is overkill for a lot of stuff.
Increasingly, people are doing things in Python. I didn’t like Python in the beginning because I didn’t like how it forced you to use indenting a certain way, but I’ve warmed up to it (though I am a bit out of practice).
In my work, I use R the most, though my firm wants to migrate to MATLAB eventually.
You can’t ask this question without letting us know what your opportunity cost is. Learning Matlab as opposed to sitting on your couch and playing video games? Yes.
Learn MATLAB, VBA, SQL, and R
Thanks for nice advice.
Can you please suggest in what order?
Thanks sir for useful advice.I will probably go for SQL or SAS as per the job requirements.
SQL can be learned at $100 in India.I don’t know about Matlab.I will learn this languages with my job.
You wanna be a programmer or an analyst? You’re not going to be a BSD on the basis of your Java skills. Certainly you can do well with technical skills, but they don’t make you the boss. Look at guys that advance at your shop and others. Its likely soft skills and business acumen that drives them up, not programming skills (though a base competency is always useful). Like FT said, this is basically an opportunity cost situation. Learn some of this (there are good suggestions on what to learn here if you do) if it means you’ll spend less time daydreaming about 10 packs. But don’t waste time you could be networking or being a better analyst or leader. The most successful are not always the most skilled, but the ones that use their skills and the skills of others to add the most value.
+1.
But as a new kid in the industry…I need to know basic of programming skills to attract employers…
I wouldn’t be surprised if it is just some sort of filter that the HR people are using. I doubt you would be doing very much programming unless you’re doing a quant/developer role.
I would strongly recommend learning R over SAS. It might be that this one company uses SAS, but R skills will be more helpful at other firms. If you get the job, then you can learn SAS.
Don’t forget that R has the advantage of being FREE. You’d be surprised how many financial firms like the idea of R because you don’t have to pay much for it.
As I understand it, SAS has an advantage with enormous data sets where getting answers a few seconds faster makes a difference. But the $1000 per seat per year license puts a lot of people off. Stata is also good for that kind of thing and (last I checked) is substantially less.
So i am learning R after VBA
SAS is good with big data sets. I’ve heard that it’s really popular with the pharmaceutical industry because the FDA is familiar with it.
I remember Stata being very good at panel data and clustering and the like. However, it’s not as good as SAS for “big data”-type work. Since it doesn’t have the SQL functionality that SAS has, you’re pretty much limited to what you can load to memory. For most people that won’t be a problem, but if you’re looking at a panel of every stock in the market, then you’ll run into problems with Stata. That being said, I think it’s much easier to use for a novice than SAS. It’s pretty easy to use and has a wide range of statistical functions. There are also some add-ins that you can use to output your results in exactly the format that everyone uses in papers.
I can’t stand the programming languages in either of these. You really can’t take much of the knowledge of the languages and apply it to other languages.
I agree that the languages used in SAS (what little I remember from a tutorial I did a long time ago) and Stata aren’t very good. SPSS is also kinda crummy.
When I was last looking for statistical software, and before I knew about R, I ended up going with Stata. I remember that Stata had a more expensive multiprocessor option, and I thought that it also had added ways of paging memory so that it didn’t all have to be loaded in RAM. I could be wrong, since I do it all in R these days (which I don’t think can handle tick-level data sets, even slowly, although there are probably some packages that also use paging to handle memory).
So far, R has managed to handle everything that I’ve thrown at it, but since I mostly do stuff with indexes, commodities, and currencies and I don’t have to deal with intraday data, I’m not that surprised. If I were doing stuff on the Russell 3k using intraday data, perhaps that would be far trickier.