|
As user start using DW, the application widens and people start looking for-
- Drill down to the transaction level (for example if sales performance of a certain sales office is in red, which sales staff have not been performing) for a better root cause analysis.
- Transactional level Reporting- As mentioned in why data warehouse, you will DW a good reference for generating your enterprise transaction reports.
- Due to generally better and more sanitized data quality in data warehouse and also it being an offline platform, DW becomes a good source for enterprise reporting.
There are other more fundamental reasons for this tip:
- Single reference point for all information needs: If you use DW as a single point reference, it will eliminate lot of reconciliation needs. This single reference point is possible, only when you have detailed data in DW.
- Auditability of Data: Data in a DW may go through multiple stages of transformation. If you want to have the auditability of data. Detailed data provides a smooth audit trail, instead of summary data, whereby different records within that summary data could be having different audit trails.
- Maximize the benefit of your ETL efforts: You would have done your extraction and transformation at the level of detailed data, before aggregating it for Data Warehouse. Therefore, if you have processed the detailed data already, it will be relatively a lesser effort to load the detailed data in DW. In other words, a good proportion of ETL efforts will remain the same irrespective of the level of detail you store in the DW.
- Future Applications- While in the beginning you may be using the DW for purely summary data analytics, over time you will find many more applications as users taste the blood. Most of these applications (like enterprise reporting, data mining...) will be requiring data at detailed level. Prudent data warehouse architecture will include this future scenario in the scope and sizing of the DW.
- Processing and storage capacity is less of a constraint now as it was few years back: This does not mean that infrastructure is not a challenge with DW. The size of data is also growing. However, we feel that cost per unit of storage and processing has fallen significantly. Moreover new technologies are emerging, which can handle data more smartly (like handling data explosion)
|
