KEDA - Data Warehousing & Business Intelligence Management

Data Mining / Predictive Analysis

According to Gartner, enterprise data collection doubles each year, yet enterprise personnel have less knowledge about the data that is collected. In a 2005 Deloitte Consulting survey, 80 percent of respondents indicated their data quality was insufficient for use in forward-looking management planning and strategy formulation. Data must be available across agency boundaries to allow executive level management to perform Predictive Analysis. This analysis can only be done if an unprecedented depth and breadth of data are available for rapid query to currently unknown opportunities.

The Commonwealth currently has an initiative underway with several state agencies to review their current data warehouse and business intelligence capabilities and deficiencies. This information will be compiled to create a ‘current state of DW/BI’ in the Commonwealth. This will be compared to the desired ‘future state’ to allow the creation of a ‘Roadmap’ document that will outline specific tactical projects to complete that can move us forward toward that desired future state.

View presentations related to the Enterprise DW/BI initiative at:  http://gotsource.ky.gov/docushare/dsweb/View/Collection-52409.

OLAP Analysis

Online Analytical Processing (OLAP) by definition should be performed against data that is stored separately from the Online Transactional Processing (OLTP) data entry system, when it is possible and cost effective to do so. This will provide multi-dimensional inquiry capabilities (‘slice and dice’) and quick response time while avoiding interference with the day-to-day operation of the data input system. To additionally reduce data input delays, predefined reports that provide tactical information may be run against the OLTP system, if possible, during non-transactional hours.

Ad Hoc Query & Reporting

OLAP data warehouses should be designed to allow ad hoc reporting. However, the ability to run ad hoc queries should be limited to a relatively small subset of knowledgeable end users who are extremely familiar with the data. The agency Data Stewards should be included within this subset. Ad hoc reporting should not require technical expertise with a particular programming language or SQL; rather it should be initiated from an intuitive ‘drop and drag’ or ‘point and click’ Graphical User Interface (GUI).

Enterprise Reporting

A Business Intelligence Competency Center (BICC) is an essential step to enterprise informational success. A BICC will leverage limited staff with specialized skills to align with business users and deliver a proper BI strategy. This approach is more effective than deploying teams for each individual project request. The more efficiency state government can show in linking data across systems and using it for proactive decision-making, the faster it can reduce costs and provide more informed decisions for the benefits of its constituents state-wide.