Data Warehouse Design - Chapter 1.7

Question 1

What is the last level common to all datawarehouse architecture types?

Answer 1

Analysis of data

Question 2

What are the 3 best approaches for end-users to query data warehouses?

Answer 2

1. Reports
2. OLAP
3. Dashboards

Question 3

For which users are reports useful?

Answer 3

Users who need to have regular access to the information in an almost static way.

Question 4

What is a report?

Answer 4

It is a query or serie of queries on data that is often frozen. There is often a fixed layout, which can take the form of a table or a chart.

Question 5

In what two ways are data warehouses beneficial to reports?

Answer 5

Reports have been around forever. But datawarehouses:

1. They contain consistent and integrated data
2. They seperate transactional and analytical analysis - improving performance.

Question 6

What is OLAP?

Answer 6

Online Analytical Processing.

Using tools (with an easy GUI), it allows users to analyze multidimensional data from multiple perspectives. This is often done through the use of several queries, and the data is often displayed in tables (because multidimensional data is difficult to comprehend)

Question 7

What does an OLAP session consist of?

Answer 7

Navigation path - A sequence of queries - The result of the queries.

Question 8

What is the navigation path in OLAP?

Answer 8

An analysis process for facts acording to different viewpoints and at different levels of detail.
This path is turned into a sequence of (sequential) queries, each referencing their previous queries.

Question 9

What are the 6 most common OLAP operators?

Answer 9

1. Roll-up
2. Drill-down
3. Slice-and-dice
4. Pivot
5. Drill Across
6. Drill through

Question 10

What is the roll-up operator?

Answer 10

It causes an increase in data aggregation and removes a detail level from the hierarchy. Example: From data every month to data every quarter.
You can also roll-up by complelety removing a characteristics dimension. Example if you normally show revenue per quarter per product group per customer per region - and you remove customer.

Question 11

What is the drill-down operator?

Answer 11

It reduces data aggregation and increases detail. Example: If you have Revenue per quarter per customer region and you then subdivide the region into all the cities in that region.

Question 12

What is the slice-and-dice operator?

Answer 12

Slicing: Reducing the number of cube dimensions after setting one dimension to a specific value.

Dicing: Reducing the set of data analyzed by a selection criterion. (Setting more than one specific values ?)

Question 13

What is the pivot operator?

Answer 13

Rotating the cube so you can look at the data from a view different perspectives. (Mainly formatting ?)

Question 14

What is the drill-across operator?

Answer 14

Creating a link between two or more inter-related cubes in order to compare their data.

Question 15

What is the drill-through operation?

Answer 15

Switching from multi-dimensional aggregate data in data marts to operational data in sources or the reconciled layer.

Question 16

What is semi-static reporting?

Answer 16

Intermediate approach between static reporting and OLAP.

It gives the user some margin of freedom. They can follow a limited set of navigation paths (queries).

Question 17

What are the advantages of semi-static reporting?

Answer 17

1. Users require less skill to operate them
2. Avoid the risk of getting inconsistent analysis due to misuse of queries.
3. Due to the constraint you prevent inexperienced users from slowing down the system.

Question 18

What is a dashboard?

Answer 18

A GUI that displays a limited amount of data in a brief and easy-to-read format. Provides real-time data.
It is a sophisticated add-on to data warehouse systems, but not their primary goal.