The decision-support chain
The path of a figure from the operation to the decision-maker: sources, extraction and transformation, warehouse, data marts, reporting. What sets transaction processing apart from analysis, and who uses what.
- 27 cards
- 5 topics
- Free
The outline, topic by topic
Each topic is weighted the way the exam weighs it.
- Concepts
- Questions
- Sort and link
- On the diagram
- Scenarios
A few cards from the deck
Business intelligence: what is it?
Show the answer
The tools and methods that turn an organisation's data into analyses for the people who make decisions.
It starts from the data of the management applications and ends in a report or dashboard: it is the whole chain, not only the last screen.
Why is the quarterly analysis not run directly on the shop's production database?
- A massive read slows down current sales, and the history is not kept there
- A production database cannot calculate sums or averages over several rows
- A production database's figures are estimates, less reliable than a warehouse's
- Access rights stop any production database from answering more than one user
Show the answer
A massive read slows down current sales, and the history is not kept there
A production database adds up perfectly well, and its figures are exact. But it is sized for short writes: a long read competes with orders, and it does not keep the history that analysis asks for.
A production database is built for … writes on the … state; the warehouse is built for long reads on the ….
Show the answer
A production database is built for short writes on the current state; the warehouse is built for long reads on the history.
The two needs pull against each other: what speeds up recording an order slows an analysis over three years, and vice versa. Hence two separate databases.
Data warehouse: what is it?
Show the answer
A database kept apart from the applications, fed from several sources, which keeps the organisation's history in a form ready for analysis.
It is mostly read, never edited by hand, and holds years of history, whereas an application's database only holds the current state.