Building the Data Warehouse

Скачать в pdf «Building the Data Warehouse»

For some sorts of events, event mapping is the only way to measure the results. Some events and activities cannot be measured directly and have to be

Jan Feb Mar Apr Figure 7.12 Simple trends.

May Jun

Jul Aug Sep Oct Nov

new incentive plan

Figure 7.13 Mapping events against a trend line.

measured in a correlative fashion. Cost justification and actual cost benefit cannot be measured any other way for some types of events.

Misleading conclusions can be drawn, though, by looking at correlative information. It often helps to look at more than one set of trends that relate to the events at hand. As an example, Figure 7.14 shows that corporate revenues are matched against the consumer confidence index to produce a diagram packed with even more perspective. Looking at the figure shown, the executive can make up his or her own mind whether events that have been mapped have shaped sales.

consumer spending index—published monthly by the Bureau of Statistics

new incentive plan

Figure 7.14 Superimposing another trend analysis over the existing one to gain another perspective.

The data warehouse can store both the internally generated revenue numbers and the externally generated consumer confidence numbers.

Detailed Data and EIS

Just how much detailed data do you need to run your EIS/DSS environment? One school of thought says that you need as much detail as possible. By storing as much data as possible, you can do any kind of analysis that might happen along. Because the nature of DSS is delving into the unknown, who knows what detail you will need? To be on the safe side, you’d better keep all the detailed data you can get your hands on. Furthermore, the more historical detailed data you can get, the better, because you never can tell how far back you need to go to do a given DSS analysis.

Скачать в pdf «Building the Data Warehouse»