Information Technology Reference
In-Depth Information
and need improvement. Direct observation requires a certain amount of skill. The observer
must be able to see what is really happening and not be influenced by attitudes or feelings.
This approach can reveal important problems and opportunities that would be difficult to
obtain using other data collection methods. An example would be observing the work pro-
cedures, reports, and computer screens associated with an accounts payable system being
considered for replacement.
Direct observation is a method of
data collection. One or more
members of the analysis team
directly observe the existing system
in action.
(Source: © Kriss Russell /
iStockphoto.)
When many data sources are spread over a wide geographic area,
questionnaires
might
be the best method. Like interviews, questionnaires can be either structured or unstructured.
In most cases, a pilot study is conducted to fine-tune the questionnaire. A follow-up ques-
tionnaire can also capture the opinions of those who do not respond to the original
questionnaire.
Other data collection techniques can also be employed. In some cases, telephone calls are
an excellent method. Activities can also be simulated to see how the existing system reacts.
Thus, fake sales orders, stockouts, customer complaints, and data-flow bottlenecks can be
created to see how the existing system responds to these situations.
Statistical sampling
,
which involves taking a random sample of data, is another technique. For example, suppose
that you want to collect data that describes 10,000 sales orders received over the last few
years. Because it is too time consuming to analyze each of the sales orders, you can collect a
random sample of 100 to 200 sales orders from the entire batch. You can assume that the
characteristics of this sample apply to all 10,000 orders.
questionnaires
A method of gathering data when the
data sources are spread over a wide
geographic area.
statistical sampling
Selecting a random sample of data
and applying the characteristics of
the sample to the whole group.
Data Analysis
The data collected in its raw form is usually not adequate to determine the effectiveness
of the existing system or the requirements for the new system. The next step is to manipulate
the collected data so that the development team members who are participating in systems
analysis can use the data. This manipulation is called
data analysis
. Data and activity mod-
eling and using data-flow diagrams and entity-relationship diagrams are useful during data
analysis to show data flows and the relationships among various objects, associations, and
activities. Other common tools and techniques for data analysis include application
flowcharts, grid charts, CASE tools, and the object-oriented approach.
data analysis
The manipulation of collected data
so that the development team
members who are participating in
systems analysis can use the data.
Data Modeling
Data modeling, first introduced in Chapter 5, is a commonly accepted approach to modeling
organizational objects and associations that employ both text and graphics. How data mod-
eling is employed, however, is governed by the specific systems development methodology.
Data modeling is most often accomplished through the use of entity-relationship (ER)
diagrams. Recall from Chapter 5 that an entity is a generalized representation of an object
type—such as a class of people (employee), events (sales), things (desks), or places (city)—
and that entities possess certain attributes. Objects can be related to other objects in many
