Databases Reference
In-Depth Information
to jot down any ideas that you have now, so that you remember to share
them during the session. Then, take this one step further and allow yourself to
dream. What would a perfect world look like to help you make decisions?
Too often, the first thing that comes to mind is to have all of the data
available as soon as possible. Of course, because there is a cost associated
with preparing, loading, and maintaining data, this is not feasible. Therefore,
consider what data would help you the most. Again, try not to focus on what is
wrong with the data today, but think about what you are trying to accomplish.
Then let the team track down and address problems with the data.
What Else Would Help the Project Team?
As you participate in the requirements gathering process, it would be helpful
for you to understand what the project team is hoping to glean from these
sessions. As discussed earlier, there is a need to understand the business itself,
what functions are performed, and how information is used today. The team is
also looking for specific clues that will help drive the next steps of the project.
The discussions will highlight the areas that are most important to the busi-
ness and identify problem areas that need more research. The project team is
not expecting to find all of the answers up front, simply to identify who and
wheretogotonext.
There are also other specific things that the project team will be looking
for. The team will sort through all that is shared to distill the essence of what
you really need. Remember that the team represents both the business and the
technical perspectives. Many clues are provided during an interview session
that shed light on how the data should be organized to support business
needs. These are not communicated in data modeling terms, but the details
are mentioned. For example, an interviewee may state, ''I need to track sales
performance by employee, by product line, by individual product, by month
and week, by sales division, and by customer state.''
Two different things can be gleaned from such a statement. First, the
business needs to track sales performance. Further questions may determine
that sales performance is measured by unit sales, dollar sales, and unit sales
as a percentage of quota. The project team should recognize these as part
of a dimensional model, called facts. Every word mentioned after ''by'' is
a characteristic that is used for reporting. In this case, there is a need for
employee, product, customer, and date data. Second, for products, the specific
attributes of product line and product itself were mentioned. These are used to
group the sales performance measures and/or to select a subset of data to look
at. These are also clues to help design the dimensional model. The project team
should recognize these characteristics as belonging in a dimension. Chapter 7
provides details about dimensional modeling.
