![](bg3.png)
3.4 Timeliness or Maturity
There is no point in having a list of clients or
products, if that information is one week old already.
Commercial data, especially in finance services, are
changing very fast and sometimes very significantly.
Some authors describe and some vendors say that
they provide ‘real-time’ Data Warehousing
solutions. We have to be very careful with definition
of ‘real-time’ and we would like to use a term ‘right-
time’ instead (Connor 2003).
3.5 End-User Involvement
Any company has top experts where key knowledge,
competence and experience can be found. The
company’s Data Warehouse and its content should
be reviewed very carefully under the guidance of
these people. The representative of the Data
Warehouse project (usually an IT person with good
business understanding) taking part in those
discussions should be an expert of the same level –
otherwise it could lead to the situation, when
business experts influence the process, outside their
expertise (even, if they believe they are IT experts
too). If it’s not the same persons, real end-users
obligatory should be involved in that project stages,
where maturity, end-user application features and
related issues are discussed.
3.6 Source and Type of Measures
Especially specific for financial institutions is a very
high number of measurements to analyze, majority
of them are derivative (like client profitability,
average turnovers)So, instead of putting attention on
data mapping attention to data transformation,
aggregation and interpretation should be paid.
Financial industry’s company may have a
comparatively high number of source systems
because of the wide range of their business activities
and because of the interest in the key measurements
of any other industry or common knowledge bases.
It’s all related to business opportunities and threats.
It’s very common, when one particular or set of
some business areas build separate Data Marts
Although there may be a lot of source systems,
they usually are similar. Those, which are related to
the business processes, are based on the granularity
of business object (client, product) and many
different dimensions. Those, which are related to the
external statistic information, usually are provided
by some state institutions or similar for-profit
institutions and are pretty similar by the content.
4 HIGHER EDUCATION
The user-driven approach is used in the development
of the data warehouse at the University of Latvia.
Education belongs to the non-profit sector. The
University of Latvia is a higher education
establishment with 30,000 students. Its business
processes include education, research, finance and
the university management - all areas are equally
important for the university’s successful functioning.
The top management is deeply interested in
gaining objective criteria to estimate these and to
support making new decisions. The administrative
director of the university is the main sponsor for the
data warehouse project, also other IT projects in the
university are initiated by top management. It is
difficult to choose the right priorities for data
warehouse development at the university. This is not
a profit-oriented business. Therefore, the usual goal
for data analysis - the profit is not the case.
However, money matters like the study fees and
other payments have to be analyzed.
The data sources for the data warehouse are the
Student Information System and the Finance
System, two Oracle databases, but they are not
integrated. The first one is developed in the
university; the second is a commercial product. This
is typical solution for many universities
In the data warehouse project of the University
of Latvia, we applied the user-driven approach. The
approach is based on the interviews.
In our methodology we used some ideas from
Kimball and Ross (2002), namely, how to organize
interviews. These ideas we supplemented with our
ideas, how to manage and use the gathered
information.
After discussions with the main project sponsor,
we defined the groups of interviewees. The
potential user groups are the following – the top
management, the department leaders and the deans
of faculties. The last group is the users whose
responsibility is data analysis, the employees from
the departments and the administrative staff in the
faculties.
The interview content was modified for each
group of users.
In the selection of the interview questions we
followed the principle – the priority is given to the
questions that find out the business objectives and
measurements. The questions in the interviews were
divided into two groups: „Business goals and
influence factors” and „Data analysis demands”. The
following questions from the interviews could serve
as examples:
– What are the goals of your department? What do
you want to achieve?
ICEIS 2004 - DATABASES AND INFORMATION SYSTEMS INTEGRATION
554