Anyone here uses MATLAB to do financial modelling?
If not, what do you use for risk analysis?
>he fell for the MATLAB meme
I use excel and pen&paper™
>>1167257
Thanks grandpa!
>>1167261
>It's newer and more complicated, it must be better!
This is why millennials need to die.
>>1167278
confirmed for never using matlab
it is way more useful and powerful than excel you just have to be willing to learn it
>>1167230
Yes, these people don't know what they're talking about. After you know how to use it then you will be able to be much more effective than someone using pen and paper.
Whenever you get into creating a black box for quantitative analysis you'll need some matlab, R, C++ etc.
>>1167300
>After you know how to use it then you will be able to be much more effective than someone using pen and paper.
Funny.....I don't think Einstein would agree.
>>1167435
kill yourself faggot
>>1167284
I don't know.
What's something you can do in MATLAB that you can't do in excel?
>>1167230
I use it, but mainly because of my school requiring I learn it.
I had to learn it for both of my undergrad degrees and still use it in grad school.
I have to use that unless piece of shit environment for an Engineering course right now. It's basically a Fisher Price replica of an IDE like java or python, but 100x more intensive on memory, full of redundant commands and loaded to the top with absolutely dumb syntax that goes against any intuition. It's basically computer science for grandma's and nine year olds. Anyone who ever thought using it was worth their time is a laughable autist.
>>1168264
>What's something you can do in MATLAB that you can't do in excel?
Solving differential equations and doing matrix multiplications.
>>1167230
R? SAS? Python?
I studied MATLAB in college, gosh that sucked.
Subpar language, stick with real ones like python. Only hippie physics professors use matlab.
>>1167309
>MATLAB
>hipster grade
it's used by most people in EE
>>1168264
>>1168489
Looking around it seems you can do matrix multiplications in excel, just not in an easy or elegant way.
Intuitively I'd say any iterative process like EM or k-Means would be very difficult in Excel, as well as data mining methods like PCA and any kind of modeling, for instance Gaussian mixture models.
>>1167230
Also,
>If not, what do you use for risk analysis?
MATLAB would be useless for middle management voodoo like this, where you'd regularly want to massage the """data""" to get your contract through. It beats Excel wherever you'd need to do multiple stages of manipulation on the same data, because the matrices don't occupy Cartesian spreadsheet space and you can just refer to them by name.
It loses to Excel if you just need to make a big rounded Powerpoint-friendly table where some cells are red and others are green with positive business memes in them.
>>1168694
I thought Excel was just some meme program where you could only make spreadsheets and nothing more.
What do you guys think about VBA?
>>1167888
If Einstein were alive today, then he'd certainly be doing all of his worth through computer programs instead of using pen and paper.
>>1169037
depends on how fast your task should run, VBA is slow in general but can do lots and is easy to implement. But can't compete with R in terms of speed for example and I won't even talk competing with Cython/Rcpp/C++. Sometimes however, for example at a work place you are limited to just excel installed on your machine, e.g. I interned at GS and there was no Python/R etc. installed, for projects we literally pushed Excel/VBA to the limits. We had often had sheets that are >200MB in size and the VBA-code had to run for like 4 hours sometimes. So yeah, VBA is shit when you have serious amounts of data (--> use R), but VBA is easy for manageable tasks.
>>1168496
yes, R/Python can be recommended for pretty much everything, above all R for data-intensive computations (statistics/MonteCarlo/bootstrapping etc.), very powerful if you know what you're doing
>>1169037
VBA is total shit, but even knowing basic programming lets you have "ADVANCED EXCEL SKILLS" which hiring managers cream themselves over
fucking vba is the slowest buggiest piece of trash i've ever had to use, much prefer R/Python
>>1169633
>I interned at GS
How did you get that internship?
>>1167300
I agree with R but even then unless your specturm level autistic and honestly think you're going to beat a team of bank quants on your own with 'le meme stackoverflow algorithmics" you're in for a shock
>>1167278
Wow, I hope you like performing gauss-jordan elimination on 1000x1000 matrices.
>>1170872
>gauss-jordan
>not outsourcing time consuming but simple activities to pajeet & co for pennies on the dollar of what i get paid
>thinking machines are cheaper than 3rd worlders yet
>this is why we make shoes in china not usa
>>1170879
Dude what. That's one line of code in matlab.
x = A \ b
That would take days to do by hand on a 1000x1000 matrix, assuming you don't make any mistakes.
>>1168665
This, but I'd say more on the communications side of things. At least in power (both utility and power electronics) we hardly ever use MATLAB.
>>1170883
Most linear algebra operations are highly parallelizeable so there's not really a reason not to use 3rd world manpower instead of expensive and difficult software.
>>1170971
>matlab
>expensive and difficult
On the topic of Risk Analysis, I've recently read The Black Swan by Nassim Taleb and he makes frequent reference to Monte Carlo Analysis.
From the way he talks about his "Monte Carlo Machine" you'd think it was magic in that he could just plug anything in and get exact results--anyone have any experience with this? Is it actually legit?
>>1171080
>paying your sorry white ass to use it
>muh living wage
>muh stem degree pay
>not just paying pajeet and friends pennies and saying hi over skype once a month
>>1171093
First of all, never refer to Taleb as if he knows or says something of relevance. The guy is a clown. Entertaining sometimes, but still a clown.
In "general" his take on a "Monte Carlo machine" means that you essentially run an analysis multiple times for example with slightly different inputs to get different end results. You then compare how different or volatile or spread out the end results are to understand the system better.
It is actually really simple to do. It can be done in Excel (just have different columns represent different "worlds" with different inputs or random innovations). This is how some difficult to price options or instruments are analyzed because they do not have "closed form" solutions due to weird payoff structures or non normal innovations.
But it is preferably done in Matlab because the code documents the process better than the individual formulas in Excel cells. Additionally it is faster by a billion orders of magnitude. What sometimes happens is that you end up prototyping stuff in Excel (if it is simple) and then implement in Matlab. However, the Excel step quickly disappears if you have to do anything remotely interesting.
>>1170879
Kek.
>>1170957
True.
Studying EE and its mainly used for controllers and FFT. Most code I've done is to manipulate systems. I'm sure it can be used for financial modelling, but until you get very used it its going to be the biggest pain in your ass.
Butt... theres a shit load of resources and tutorials.
>>1171398
Why does everyone hate MATLAB?
>>1171859
Because this board is a bunch of braying retards who don't have real world jobs.
STATA
>>1167278
Excel was created in 1985. Matlab was created in 1984.
>younger people don't understand the world as well as I do, an old person afraid of technology
>>1170872
I've done matrix manipulations you people wouldn't believe. Determinants of sparse matrices calculated in polynomial time. I watched MATLAB find the spectral radius listening to anime trance at 2am. All those moments will be lost in time, like tears...in...rain. Time to die.
>>1171963
Why not start up MATLAB and do some good ol' matrix computations?
>>1171963
This is next level kind of feels brah.
>>1171935
Kek.
WHO C++ MASTER RACE HERE?
>>1172564
Isn't that what all the quants use?
>>1167230
I spent 7 years using Matlab to do data analysis, image processing and Monte Carlo (not in finance though). Python would probably be my choice at this point because it's free, but excel is good to play around with if you are just trying to do SQL-like things (figuring averages, counts simple threshold type strategies etc...)
>>1173071
What kind of job did you have where you used MATLAB so much?
I had to take a course on it in college even as a Civil Engineer. I personally enjoyed it. There are probably better languages out there for sure, but it does have some strong points like image processing and web scraping.
Made an awesome script that predicted outcomes of College Football games as my final project. It batted around 60%, which was pretty cool.
>>1173978
Physics graduate research.
>>1171365
and excel is slower than everything, and excel can not use multithreading or multithreading gpu.
Excel is great to manipulate fast small datasets.
So I know a little bit of R and I'm in the business of geology and resource modelling.
Any benefit for me to pick-up C++ or Python or deepen my knowledge of R?
>muh matrix
Give me one PRACTICAL example where you have to multiply matrices in everyday life
>>1176813
Geostatistical analysis.
>>1176813
Every materials related simulation (FEM,FVM,FDM...).
The everyday life of an engineer though.
>>1171859
not sure, but I hated Matlab until I realized how much easier it made my life. Its like any coding language, its a bitch till you truly understand it.
>>1177025
When was the point you finally started liking it?
I'm used to the syntax and I like the plots it makes, that's why I use it.
