Data Mining our Business

Up to this point the business analyst has contributed to the business intelligence initiative in the following ways:

1. 1 Conducted stakeholder interviews to understand the business needs
2. 2 Executed business usage scenarios to identify meters, metrics, and KPI candidates
3. 3 Assisted in the dimensional modeling exercises in preparation for OLAP development and preloading activities
4. 4 Participated in data profiling activities such as data cleansing, and achieving better quality data by way of improving business processes

I discussed dimensional modeling in the last post and why I felt this was something the BI BA can own. I also focused on red and blue people (nothing political here). These colors represent the business and IT. When blended they become “Purple People”, a Wayne Eckerson (TDWI) term. I wanted to continue down this path with data cleansing and data quality, and how BAs can assist in this area.

Data Cleansing

Data cleansing attempts to understand the content of the data, establish a standard form, and correct any inaccuracies detected. Data cleansing is done prior to the extract, transform, and loading (ETL) of the data. One of the tasks associated with data cleansing is the removal of obvious duplicates and not so obvious duplicates.

Double Agents

DW/BI BAs cannot truly be successful by functioning as required by standard SDLC projects.  We must make the transition into “Purple People”, a term coined by Wayne Eckerson of The Data Warehouse Institute (TDWI).  In a DW/BI project there are red people (technologists) and blue people (business folks).  The blending of the two results in The Color Purple.  Purple People often possess ubiquitous business and IT knowledge, has credibility, and can play the double agent role quite well between business and IT.  One way a BA on a DW/BI project can head towards purple is to augment their subject matter expertise with dimensional modeling.