NAME

Finance::Shares::Overview - Outline of Finance::Shares modules

DESCRIPTION

Overview

This suite aims to provide a toolbox for analysing prices of shares quoted on the world's stock exchanges. It gets the quotes from http://finance.yahoo.com, storing them in a mysql database. Calculations and tests may then be applied to the data, in an attempt to derive some meaning from the semi-chaos.

Input can be via a web interface Finance::Shares::CGI and the results are output as graphs Finance::Shares::Chart. The processing, supervised by Finance::Shares::Model focuses on applying tests to quotes for a company over a range of dates, giving an historical perspective. Of course, if the date range just covers today, the current position is evaluated. Mostly the tests compare results from calculations applied to the price or volume data. These functions are provided by several modules such as Finance::Shares::Averages and Finance::Shares::Bands.

Stock quotes are fetched using Finance::Shares::MySQL and held in a Finance::Shares::Sample object. Functions like averages or trend identifiers are applied to these data and the results used in tests. Each model can apply several tests to several samples. When the tests are run signals may be invoked highlighting interesting situations. The intention is to use these signals to drive a simulated portfolio which can be used in analysing risk.

Design Decisions

Charts are output as a PostScript::File. Originally this was because I couldn't find any software that printed A4 sized charts with enough accuracy and detail to scribble lines on. In practice, I don't print them out nearly as often as they are viewed on screen, but it is good to be able to do both. This doesn't sit particularly well with the web interface, but programs like pstill can convert the output to PDF format easily enough, and most browsers can be configured to do something useful with files in application/postscript MIME format.

Another initial impetus was the need to have quotes available offline, and this is still useful. However, this is not essential as the MySQL module can be configured to fetch only (not cache).

Modules for Momentum functions, Oscillator analysis and identification of Support and Resitance trends are planned or in development. These mostly follow John J Murphy's book "Technical Analysis of the Financial Markets". Also planned is a Portfolio module for holding stock positions controlled by signals generated from Models. It would be nice if that could include some risk analysis, but that's crystal ball land.

The CGI interface was added to make the plethora of options a bit more manageable and it does provide a useable (if slow) front end. However this suite was implemented in perl to provide flexible scripting tools. For example, it should be possible to automate a series of tests to quotes for several hundred companies in order to identify the stock with the most regular patterns. I haven't included support for signals in this release because I'm still not sure what form the callback function should take. So if you have any thoughts, do email me.

Tutorial

There are always two paths, aren't there?

Finance::Shares::CGI provides a graphical user interface with lots of help and explanation alongside all the options. Is it the 'easy' route? I don't know. It's certainly harder to set up than a script. But it's probably easier to explore if you prefer windows. I'm not going to describe how to use the CGI interface, but here are a few hints.

There is a 'run' button at the bottom of the Model page. You will need to set up a Sample specifying the stock and dates you're interested in. The next step would be to add a Function or two like a 7 day moving average or 30 day highest/lowest values. Adding a Test on those functions will mark buy or sell signals, depending on what you've chosen. After that, it's just a matter of seeing what the different options do.

The trouble with the CGI route is that that's all you learn. You can only explore the paths provided and you have to start learning all over again if you want to try anything original for yourself. But it does give you some idea of how the land lies.

The tutorials attempts to guide you along the more adventurous path. They cover much the same ground, but hopefully you will then feel more confident about experimenting for yourself when you have seen some scripts working. The lessons available can be listed with the command line:

    man -k Finance::Shares::Lesson

Preparation for the tutorials

You need mysql working on your system with the test database that comes as standard. The user and password used in the scripts will also be called 'test'.

As root, fire up the mysql client and declare the user:

# mysql Welcome to the MySQL monitor. Commands end with ; or \g. Your MySQL connection id is 3 to server version: 3.23.52-Max-log

  Type 'help;' or '\h' for help. Type '\c' to clear the buffer.

mysql> grant all privileges on test.* to 'test' -> identified by 'test';

  Query OK, 0 rows affected (0.15 sec)

NOTE: There is always a database called 'test', so you don't need to create it. However, if you've chosen to use a different name (db_name) you will need to give the command create database db_name; as well.

All the lesson scripts are found in the 'tutorial' directory of the Finance::Shares package. Exercises that are a little more challenging are marked with an asterisk (*).