Simple Complexity

Open Source Business Intelligence

July 5th, 2006 | by Neal Levene |

In DM Review Extended Edition June 2006 (click here to download, the article is at page 8), Steve Miller has an excellent article about the use of open source technology in for analytic intelligence.

Miller discusses a Python, PostgreSQL, OpenOffice, and R BI platform.

Python is nominated for the tool of choice for the text and data processing steps that are at the core of ETL.
PostgreSQL is recommended as the database for the data warehouse.
OpenOffice is suggested to handle query and reporting as well as what-if analysis. OpenOffice modules include Writer (word processing), Impress (presentations), Calc (spreadsheet), Base (database), Draw and Math. Base offers GUI-based query construction. The results can be shared with Calc for what-if analysis. Base forms can be used to develop simple front ends tied to database tables with little effort.
R’s role is analytics, visualization and predictive modeling.

From an integration perspective, Miller suggests:

Python, the scripting language, can manage the data and system administration tasks, kicking off and conditionally executing jobs, assuring that tasks are completed as scheduled. Python, the database language, can access PostgreSQL efficiently through DBI and ODBC that, in tandem with Python’s file, string, XML and URL modules, provide a foundation for ETL. OpenOffice provides the window to the database. R, the data analysis language, is used to transform data to intelligence. . . . Python, the agile language, functions as the glue to unite all tools. The Python/R community has developed RPy, a module that allows access to R functions and data structures from within Python programs. With RPy, all of the capabilities of R - functions, models, graphics, etc. - are available to Python data structures that can, in turn, write to R. This allows Python to coordinate the administrative, database, analytical and Web-delivered aspects of BI, fundamentally enabling complete BI solutions in Python.

The whole article is good. Check it out here.

I’d love to hear about your experiences with open source business intelligence tools.