I'm looking for someone to write a market analysis program that creates sets of recursive databases from stock price calculations. The program should run on a Linux (Ubuntu) system.
Overall, the program would perform the following steps:
1) At specified time intervals, download a set of stock quotes from a given web address in .cvs file format.
2) Develop a set of historical prices by recording the price and time of each quote in a unique database created for each separate stock (Database A for Stock A).
3) Perform a given calculation on adjacent prices in the historical databases (subtracting, dividing, etc) and store the results and time in a new database (Database B for Stock A).
4) Perform the same calculation on the newly created Database B to create a subsequent database (Database C for Stock A). Continue to perform the calculation on each new database for a number of times specified by the user (creating Databases D, E, F, etc for Stocks A, B, C, etc).
5) Take adjacent values in each database and create strings of digits of a given size. Store the sequence itself, the number of times it occurred, and when it occurred in another separate database.
6) Determine if the most recent string of stock quotes from Database A matches any of the string occurrences and, using the set of historical prices, create a probability that the stock will both rise and fall in a given number of days and to what degree.
7) Output the overall top occurrences and their probabilities to a website and update at specified time intervals. I'd like to be able to view the final data through a web browser, with the occurrence strings displayed as charts (resembling stock charts), the probabilities and high/low values as digits, and the ability to click on an occurrence to display the top probabilities for that particular stock.
The following values should be adjustable by the user:
1) Time interval to get stock quotes
2) The calculation performed
3) The number of times the calculation is performed (ie the number of databases created for each stock)
4) The degree of precision for the calculation results (hundredths, thousands, etc)
5) Number of digits in a string.
Finally, I have a set of historical prices for each quote in .cvs file format that i would like to be able to add to the each databases of historical prices (Database A).