DATA MANAGEMENT

Data management is one of the important tasks for an IT department of any organisation since IT applications cannot be done without having the right data.

However, managing data is increasingly difficult due to the following factors:

 Most organisations require past data to be kept, while new data need to be added rapidly. Thus, the amount of data required increases exponentially with time.

 Most organisations also require an ever-increasing amount of external data.

 Data in most organisations are not stored centrally. They are collected by many individuals and departments and stored at various locations. Some of the data are not even available in digital format.

Like other things, data can get obsolete: new data need to be collected and old data need to be updated. Thus, data management can be visualised in the form of data life cycle as shown in Figure 7.1.

Figure 7.1: Data Life Cycle

7.1.1 Data Collection

Data collection is a process of collecting data from the data source. There are three different types of data source:

(a) Internal Data Source;

(b) External Data Source; and (c) Personal Data.

Internal data source is located within the organisation and normally captured by the organisationÊs information systems and stored in the database or physical files. Examples of internal data source for a university are:

Ć Data about students;

Ć Data about lecturers;

Ć Data about courses; and Ć Data about facilities.

Examples of external data sources are:

Ć Data about other organisations, especially our competitors. This data can be obtained from the organisationsÊ websites, annual reports and published brochures;

Ć Government reports, for example annual budget and economic reports; and Ć Data about other countries.

approaches for implementing given tasks or solving given problems need to be captured and stored, so that they can be referred to by other employees in the future.

7.1.2 Data Storage

Most of the information systems store data inside databases. A database is a collection of data organised to serve one or more application systems. The management of data inside a database is done by a software called Database Management System (DBMS). DBMS acts as an interface between application systems and the database as shown in Figure 7.2.

Figure 7.2: Role of a DBMS

There are different types of DBMS, for example Relational DBMS, Hierarchical DBMS and Object-oriented DBMS. Relational DBMS is the most popular type of DBMS. In relational DBMS, data are stored by using two dimensional tables. For example, data about students, lecturers and courses may be stored as tables shown in Figure 7.3.

Metric No Name Address Programs A45167 Ali bin Abu Kuantan BIT A46715 Ramasami a/l

Muthusami

Staff No Name Qualification Specialisation K003478 Syarifah Aminah PhD Computing K003479 Tan Sook Meng MSc Accounting

K003678 Ahmad Harun MIT Computing

Data TMK1213 Programming 1 None TMK1223 Programming 2 TMK1213 TMK2243 Web-based Systems None Data

about Courses

Figure 7.3: Tables describing data about students, lecturers and courses

The current trend in data storage is to use a concept called data warehouse. A data warehouse is a repository of all of the data needed by the an organisation.

Some of the characteristics of data warehouse are:

Ć Data are stored by subjects (for example by students, staff, vendors, products);

Ć All of the data in the data warehouse are stored by using similar method of coding;

Ć Data are kept for a long time so that they can be used for forecasting and comparisons; and

Ć Data are stored in a multi-dimensional structure so that data can be viewed and analysed from different perspectives.

The process of developing a data warehouse is done by extracting data from all possible data sources as shown in Figure 7.4. Since the data from the data sources are in different formats, there is a need for the data to be transformed into the standard format used by the data warehouse.

Figure 7.4: Developing a data warehouse

Although the concept of data warehouse is very good, the process of building and maintaining a data warehouse is very expensive. As an alternative, many smaller organisations build a lower cost, scaled-down version of a data warehouse, called a data mart. The advantage of data marts include:

Ć Low cost;

Ć Shorter implementation time; and Ć Easier to maintain.

7.1.3 Data Analysis

Once the data are stored inside the data warehouse or data marts, they can be analysed to produce meaningful information.

Data mining is one of the most important techniques for data analysis that focuses on modelling and knowledge discovery. Given a data warehouse of sufficient size and quality, data mining technique can be used to:

Ć Automatically predict trends and behaviours; and Ć Automatically discover previously unknown patterns.

Data mining technique is now being used by many companies worldwide. Some of the examples are:

Ć Large supermarket are using this technique to predict sales and then to determine correct inventory levels and distribution schedules among outlets;

Ć Banks are using this technique to forecast levels of bad loans; and Ć Insurances companies, for forecasting claim amounts.

The types of information that can be produced through data mining is shown in Table 7.1.

Table 7.1: Various capabilities of data mining

Capabilities Explanation Examples

Association To link occurences to single event.

A study of supermarket

purchasing pattern may find that when flour is puchased, sugar is also purchased 60% of the time.

In sequence To link events over time.

A study may find that if a person puchases a refrigrator, he will puchase a tv set within a month at 65% of the time.

Classification

To recognise patterns that describe the group to which the item belongs.

To discover characteristics of students who might not be able to complete their studies.

Clustering To discover different grouping within data.

To partition students based on their programming ability.

Forecasting

Uses series of existing data to forecast what other values will be.

To estimate future value of sales

7.1.4 Data Visualisation

Human beings find it easier to understand information if it is represented graphically. For example, it is easier for you to describe ways of going to a particular place by using a map than to descibe it verbally. Similarly, human beings can understand data better if it is represented graphically. The technique for representing data in this way is known as data visualisation.

An example of data visualisation tool is Graphical Information System (GIS).

It is basically a computer-based system for capturing, checking, integrating, manipulating and displaying digitised maps. An example of GIS systems is GoogleEarth, which enables us to view the map of the whole world. This

Figure 7.5: Google Earth main user interface

Another type of data visualisation is the virtual reality (VR), which enables people to share and interact in the same artificial environment. Some examples of VR include

 Training;

 Virtual Prototyping;

 Virtual Museums;

 Three dimensional games;

 Real estate presentation and evaluation; and

 Virtual physics lab.

7.1.5 Modelling and Simulation

Modelling and Simulation is a technique that can be used for developing a level of understanding of the interaction of a system. A model is a simplified representation of a system in order to promote understanding of the real system.

A simulation is the manipulation of a model in such a way that it enables one to perceive the behaviour of the system.

Visual Interactive Modelling (VIM) is an application that allows user to view impact of different management or operational decisions graphically. One of the widely used VIMs is Visual Interactive Simulation (VIS) that enables users to watch the progress of the simulation model in an animated form. One of the VISs is called Witness produced by Saker Solutions. Witness is a comprehensive discrete event and continuous process simulator. Another example is MapleSim, a product of MapleSoft.

7.1.6 Business Intelligence

Another interesting application is called business intelligence (BI).

Business intelligence software tools enable data to be analysed in order to produce reports, predictions and alerts. These tools can also display information in graphical presentations. The process of BI is shown in Figure 7.5.

Figure 7.5: Business Intelligence Process

One of the vendors for BI is Oracle. Oracle describes its BI application as „a portfolio of technology and applications that provides the industry's first integrated, end-to-end Enterprise Performance Management System, including BI foundation and tools ă integrated array of query, reporting, analysis, alerting,

Ć Oracle BI Warehouse Builder;

Ć Oracle Data Mining;

Ć Oracle BI Bean; and Ć Oracle Reports.

In document Fundamentals of ICT (Pldal 154-162)