Technology Blog

The key periods are the important reporting and financial periods. Does the business have special processing each week, each month, each quarter and so on? Some businesses are not interested in information in the short term, such as daily, and will tend to analyze data over weekly or larger periods. This may be because a given day’s business takes several days to process through the operational systems, with the data becoming available to the data warehouse over the period of a week or longer. It could be that the business does not do daily analysis because it never had the processing power to do it. You need to understand the reasons and the requirements. Just because they do not analyze daily data now, possibly because they have not been able to, it does not mean they will never want to analyze data on a daily basis.

Key business dates need to be documented. The start of the company’s business year, the start of the local tax year and so on are key dates, and will have effects on the processing load on the data warehouse. Some companies have standard business holidays when the business closes down for a number of days or weeks.

One of the key steps in the analysis stage is ascertaining the user query requirements. Before the data warehouse layout can be designed you need a clear understanding of the queries that are likely to be run. Although there will probably be some canned queries or standard reports that can be identified, it is unlikely that the users will have a clear definition of everything they want.
Even if the users are unusually forthcoming, and can give chapter and verse on their requirements, it is unlikely that this will be the whole story. There may be other users who are to be added later, and there is no guarantee that their queries will be the same as those of the current users.

As the users get used to the data warehouse and its abilities, they may begin to explore data in different ways. Remember, it is the ad hoc nature of a data warehouse that makes it what it is. As the users’ requirements change over time, the data warehouse will evolve, but the database design needs to remain constant. So, when performing the analysis requirements capture, it is vital to gain an understanding of the business, and the business sector it occupies. This is the one constant: even if the business changes, it is unlikely that It will change so radically that it moves business sector. A telco is not likely to become a bank, and an airline is unlikely to become a retail chain. In the unlikely event that such a change does occur, the data warehouse will probably need to be designed again from scratch in any case.