Tools used in analytics

Having learnt statistics and operations research in college days,, I have started to explore analytics for application development. While in college, the faculty provided a a small problem data , I learn that I need to construct the data, generally large in real life. While in college. faculty constructed/designed operation research or statistics/probability scenario and provided(some time copied from a book) the problem for us to solve, in industry, I need to understand the business problem and formulate the problem and then arrive with solution.

It is quite possible that arrived solutions might not be feasible for  implementation  due to people constraints or data cannot be collected due to regulatory  reasons, leading to modifying  the problem definition jointly with business owner and also updating the solution. Good news is that tools are available to  perform complex calculations and come with algorithms implemented and tested already.  The new challenge is to identify the algorithm or technique in the context of  the current problem and to supply  input data for the  algorithm or procedure defined by the tool.   Exploring how to perform application development that leverage the tool.

  • [Free] R is a free software environment for statistical computing and graphics. It compiles and runs on a wide variety of UNIX platforms, Windows and MacOS.  R is available as Free Software under the terms of the Free Software Foundation‘s GNU General Public License in source code form.
  • [Paid] AmiBroker is a  trading system and development platform, with two primary modes of operation – charting and formula evaluation. AmiBroker is a technical analysis tool to manage your stock portfolio and gives you real time quotes taking the information from on-line services. In its charting mode, historical price and volume data are displayed along with technical indicators to help analyst to look  for patterns and conditions. In the formula evaluation mode,  patterns, conditions, and rules are described using a programming language. The program analyses the price and volume data and reports on the profitability of the rules. When profitable trading systems have been found, it scans the group of stocks that are of interest to the trader and lists the current buy and sell signals. The signals can optionally be send directly to a broker for execution. Any system that runs Windows efficiently will run AmiBroker efficiently.

Explored R and Amibroker and learnt that both tools are  extensible and useful in the  respective scenarios.

R Project for Statistical Computing

  1. One needs to spend time learning R  in an interactive environment using RStudio and the tutorials at Quick-R and Learn-R got me going very fast.
  2. Python has packages that allow you to do most of the things that you want to do in R, from data wrangling to plotting. One can access R objects from Python. Check Stack Overflow  link for more information.
  3. PHP be can be leveraged to work with R.  Check GitHub link  ability to run R scripts from PHP and   How to integrate and execute with R from PHP 
  4. Vim-R-plugin : Plugin to work with R

Amibroker.

  1. AmiBroker can be extended writing new scripts. AFL code is used to represent both user interface and calculations.
  2. Programmers can package the AFL scripts to work as part of DLL library and the DLL library can be deployed in to existing AmiBroker client environment.There is Net SDK that can be used to help in developing DLLs that would work integrated with AmbiBroker. More details at link. There are plugins to convert AFL scripts to C++ plug-ins.
  3. AFL scripts can be easily and quickly converted to .NET plugins leaving no public trading logic in AFL files. .NET plugins can be protected and licensed. The protected plugins can be made public and  only customers with license for their machines can use the protected .NET Plug-Ins.