Here's my starter for 10.
To do any form of data mining on forex data, we need need:
1) Accurate historical data, ideally tick. There is of course Metatrader data from various companies, including Alpari. A source I have used in the past is Gain Capital
2) A store for the data. We could of course just dump the data into Excel, and get surprisingly far. More realistically, to do any serious stuff, we need a SQL database. This is where things get interesting. We either hook up to the database through a programming language, OR we have to have powerful built-in procedural extensions, in which case it is perfectly feasible to manage mining data that way.
I think for us amateurs, who don't want to fork out megabucks, options are:
- MS Access: complete (if picky) SQL and good programming support (VBA), however, in my opinion Access is just too non-enterprise.
- MySQL (open source): free, fast and no size limit. However, SQL limited outside of standard operations and no Analytics which is a real pain. Stored proc language is far too limited to do anything serious. Good choice if willing to hook up and program via your fav programming language.
- SQL Server Express: comprehensive SQL + T-SQL procedural, powerful, but size limited.
- Oracle Xpress: comprehensive SQL (with extensive Analytics) + PL/SQL procedural + extensive and very powerful built-in packages, but sadly, size limited.
- Postgress (open source): I have never used this, but have heard that the DB is almost as stable as MySQL, and has a PL/SQL like programming language. Probably worth investigating.
For what its worth, I use Oracle Xpress (because I love the Analytics + the built-in packages) or I go to MySQL if I hit size limits.
What's the DB of choice for others doing forex data analysis?