Exam 9: Business Intelligence Systems

Correct Answer:

Verified

Basic reports and simple OLAP analyses can be made directly from operational data. However, operational data may be unsuited for more sophisticated analyses, particularly data- mining because there are missing values or inconsistencies with the data. This is why most large companies extract and clean the data and put it into a separate data warehouse. Some dirty data may have illogical entries for gender, phone number, age, etc. Purchased data often has missing values or values that are out of an allowed range. More difficult, is when the data becomes inconsistent over time, because it does not reflect changes in the real world, such as new zip codes, area codes, addresses, income levels, etc. For a data- mining application, a few erroneous data points are worse than no data at all, because it will introduce bias into the analysis. The level of granularity can also be an issue. In other words, the analysis can only go to the level of detail at which the data was originally collected. So if a company like Hershey's want to track sales of candy, they must have to decide what the granularity of the data will be for each item; i.e. will it be by the case, the box, or the individual candy bar. A data warehouse is a separate facility where the data is extracted, cleaned, and integrated into a single format so that it can support these more complex data- mining applications.