DATA WAREHOUSE PROJECT LIFE CYCLE – Ongoing maintenance and changing user requirements

Ongoing maintenance is the final phase of creation of a data warehousing system. The ongoing maintenance of data warehouse generally involves constant loading of new data and addressing the changing analysis requirements of end users Now as the end users gains confidence, satisfaction and control over the data warehousing system, they wish to switch over to various levels of the basic data warehouse system as follows-

  • Basic analysis- is the most basic level of a data warehouse where the end user may be using the warehouse for calculation of averages and sums across salient subject areas and use of other heuristic data analysis methods and report generation.
  • Correlation analysis- Now the users begin to see the value of their data warehouse and they develop models for correlating facts across data dimensions. This stage marks the beginning of stochastic data analysis.
  • Multivariate data analysis- End users begin to perform correlations on groups of related facts and they become more sophisticated in their use of analytical statistics.
  • Forecasting- End users begin to use statistical packages to generate forecasts from their data warehouse.
  • Modelling & simulation- End users begin to recognize that they can test hypothesis against their data warehouse data and they begin to construct simple what if scenarios. Since end users are now intimate with their data, they begin to construct sophisticated simulation models. This is the phase where previously unknown data relationships are often discovered.
  • Data mining- End users begin to extract aggregates from their data warehouse and feed them into neural network programs to discover unobtrusive correlations i e. the basic approach to Data mining should be to startup as a basic data warehousing system and go through the above stages of data warehouse maturity and thus develop into a full fledged Data mining application.