Financial Open Source projects for beginners

[Wojciech Witowski]

Hello QuantNet!

I just finished the QuantNet C++ course, and I would like to get my hands dirty. I could, of course, get one of the many 'Financial modelling with or without C++' books and start implementing everything I see, but perhaps a better idea would be to contribute to an open source project. I think (and feel free to burst my bubble) that working on an open source project would be something to put on my CV, while independent development at home - would not (though could be mentioned during an interview).

I looked at CodePlex to see if there are any financial projects that need help, but couldn't find any C++ projects. I tried QuantLib as well, but doesn't seem like there's much going on there (http://sourceforge.net/p/quantlib/bugs/), especially for a beginner. I also checked GitHub and wxWidgets, but those are not financial projects (and it is also difficult to find bugs-for-beginners there). Of course, pure C++ projects are as interesting as 'quant' projects, so I could also work on those. I work full time auditing quants, so in theory I could do pure quant projects on the job but we don't get a lot of 'development time'...

Could any one suggest a good place to start? Please share your experience with open source projects, or getting non-work hands-on experience in C++/quant programming.

 

[Daniel Duffy]

First of all, congratulations on achieving the QN C++ certificate!

In general, it is important to keep C++ level on par as it is easy to forget if you leave it for a few months.

I am not sure what is best approach in the current context. IMHO more on Boost with computational finance might be a good idea.

For my own part, I have many open issues in applying the ADE method to calibrate and pricing.
http://www.datasimfinancial.com/forum/viewtopic.php?t=327

Please see my FD2 and MC1 projects here (they are hibernating but maybe they can be picked up)

http://qfcl.wilmott.com/


// Maybe time series, GARCH modelling and real-time data feeds?

 

[Wojciech Witowski]

In general, it is important to keep C++ level on par as it is easy to forget if you leave it for a few months.

I agree, therefore I'm trying to find a good project to spend my evening and weekends on. ADE sounds interesting, and just last week I read this article:

An alternating—direction implicit scheme for parabolic equations - Craig, Sneyd, 1988

After level 9 of the course I'm curious about the stability issues with the FDM, so both ADE and ADI seem like a reasonable next step.

As for FD2 and MC1 projects at Wilmott: I don't see any issues assigned to them, but I would be very interested in working on them. My user at QFCL Wilmott is WojtekW.

 

[Daniel Duffy]

Wojciech,
ADI and ADE look like very good projects to work on in C++ (you could use Boost ublas).

BTW have you seen Soviet Splitting which is a competitor of ADI?

See thesis

http://www.datasimfinancial.com/forum/viewtopic.php?t=101

 

[Wojciech Witowski]

Thank you for the link.

I've only heard about the Soviet Splitting in your lecture, but will definitely have a loot at it.

 

[Strat-O]

There are a lot of interesting open source projects out there. It seems that there were many trading-related projects being worked on prior to 2008, then many were promptly abandoned.

I downloaded the TA_lib source (which is really C, not C++) and was trying to use some of TA_lib's candlestick features and was pretty sure I was doing it right but for some reason TA_lib refused to catch any patterns. But in the process I revived some old C programming skills and relearned how to use the GNU debugger (I use Linux). Overall it was a positive experience because I was reminded how fast C was (Normally I write perl and JavaScript) and I was also impressed how fast C compiled on modern machines--explain to me why I am writing scripts again? .

So you might consider forking some simpler but abandoned project an run with it alone or maybe something similar.

 

[alain]

...
I downloaded the TA_lib source (which is really C, not C++) and was trying to use some of TA_lib's candlestick features and was pretty sure I was doing it right but for some reason TA_lib refused to catch any patterns.
...

Maybe that tells you something about Technical Analysis.

 

[Den Ryan]

I am surprised nobody mentioned Quantlib - yield curve, derivatives pricing models, various solvers...
Another is OpenGamma (http://www.opengamma.com/)
Data source? It's never "Open Source". Yahoo, Bloomberg, Morningstar everyone publishes *Some* data freely on the internet but data feed is never free. You can either code your own web scrapping tool, or buy a commercial tool such as appliedalgo.com -

You might want to check out this list
http://quant.stackexchange.com/questions/141/what-data-sources-are-available-online
https://appliedalgo.com/appliedalgoweb/doc/AppliedAlgo Application Scenario Internet Download.pdf

 

[Ali_MFE]

Good evening,

I am a UC Berkeley MFE with not much experience. I am looking to work on some open source projects close to trading or in research domain. Any suggestions? Also is someone looking for help with any ongoing project?