You are using an out of date browser. It may not display this or other websites correctly.
You should upgrade or use an alternative browser.
You should upgrade or use an alternative browser.
R for datamining/backtesting/trading
- Thread starter caementarius
- Start date
Nice presentation - it confirms what I've been doing for a few years now in terms of how to best represent/store financial data in particular full OB depth and trades data. Plain old binary files stored in a columnar manner - one file for each column be it Time, Depth Levels, or Trades. Fully portable across all platforms and applications. On top of that I use FastBit plus my own access layer to query data. Also very easy and quick to map that data into memory if needed. Never tried OneTick or KDB but any other DB I tried was miserably worse.
Quote from BMarion:
Thanks for the great R link! Where are you getting your data?
I export historical data in CSV format from TradeStation. There are pseudo-indicators and keystroke macros on the TradeStation forums that can be used to accomplish that.
I have also written API code to export data from CTS T4.
Quote from caementarius:
Or - If you've been down the same road building your own tools for the whole process, what technologies did you settle on?
For strategy dev. and backtesting I'm using Mathematica. Gives you (almost) the power and possibilities of C, but allows you to do many things in a very convenient way (due to its symbolic nature). In particular, setting up parallel computation is as simple as possible. Main drawback is the price tack, as long as you don't have access to an educational license.
Quote from comintel:
That is very reasonable.
See also
http://www.rinfinance.com/agenda/2012/talk/JeffRyan.pdf
for discussion of ways to extend R to higher frequency data.
nice. any working code samples?