Description
Business Intelligence Systems (BIS) require historical data or data collected from various
sources.
Informatica Economic? vol. 13, no. 4/2009 99
A model for Business Intelligence Systems’ Development
Adela BARA, Iuliana BOTHA, Vlad DIACONI?A, Ion LUNGU, Anda VELICANU,
Manole VELICANU
Academy of Economic Studies, Bucharest, Romania
Faculty of Economics Cybernetics, Statistics and Informatics
[email protected], [email protected], [email protected],
[email protected], [email protected], [email protected]
Often, Business Intelligence Systems (BIS) require historical data or data collected from var-
ious sources. The solution is found in data warehouses, which are the main technology used
to extract, transform, load and store data in the organizational Business Intelligence projects.
The development cycle of a data warehouse involves lots of resources, time, high costs and
above all, it is built only for some specific tasks. In this paper, we’ll present some of the as-
pects of the BI systems’ development such as: architecture, lifecycle, modeling techniques and
finally, some evaluation criteria for the system’s performance.
Keywords: BIS (Business Intelligence Systems), Data Warehouses, OLAP (On-Line Analytical
Processing), Object-Oriented Modeling
Introduction
The main objective of our research
project that we had in one of the multination-
al companies from Romania is to develop a
decision support system for public institution,
so we tried to meet the requests from the ex-
ecutives and managers of one national com-
pany. We based our work on previous expe-
riences, researches, articles and studies that
we’d been developed. Thus, we followed the
next classical steps: analyze, design, develop
and applying for the project lifecycle the
framework described in the book [2]. For the
implementation phase we used different BI
techniques, like data warehousing, OLAP,
data mining, portal and we finally succeeded
to implement the BI system’s prototype and
to validate it with the managers and execu-
tives in one national company. The system
gathers data, using the ERP system, to extract
data from different functional areas or mod-
ules such as: financials, inventory, purchase,
order management or production. For the ex-
ecutives, the system is able to provide analyt-
ical reports and dashboards. As the storage
solution we designed and build a data ware-
house. The major problem is that there will
be many more changes in the structure of the
organization and the impact of these changes
may affect the BI system. So, we need to find
a solution, and, based on our previous re-
searches on Object Oriented modeling, we
consider it as a good option. In the last years,
there have been some proposals to represent
MD properties at the conceptual level. As we
mentioned in our previous researches, based
on the conference paper [1], we defined a set
of object-oriented extensions that can be used
for modeling the components and require-
ments of a data warehouse. Also, we had to
consider the system’s development lifecycle
that has to be flexible and easy to fulfill.
2 The concept of Business Intelligence Sys-
tem
With rapid advances in technology, enter-
prises today frequently search for new ways
to establish value positions. Well built Busi-
ness Intelligence Systems (BIS) can provide
the ability to analyze business information in
order to support and improve management
decision making across a broad range of
business activities. They leverage the large
data infrastructure investment, for example,
in ERP systems made by firms, and have the
potential to realize the substantial value
locked up in a firm’s data resources [7].
While substantial business investment in BI
systems is continuing to accelerate, there is a
complete absence of a specific and rigorous
method to measure the realized business val-
ue, if any. BI systems have the potential to
1
100 Informatica Economic? vol. 13, no. 4/2009
maximize the use of information by improv-
ing the company’s capacity to structure a
large volume of information and make it ac-
cessible, thereby creating competitive advan-
tage: “competing on analytics” [15].
BI utilizes a substantial amount of collected
data during the daily operational processes,
and transforms the data into information and
knowledge to avoid the supposition and ig-
norance of the enterprises [14].
The main characteristics of a BIS are: the ca-
pability of providing representative informa-
tion to the high-level management, to support
strategic activities such as goal setting, plan-
ning and forecasting, and also tracking per-
formance, to gather, analyze, and integrate
internal and external data into dynamic pro-
files of key performance indicators. Based on
each executive’s information needs, BIS can
access both historical and real-time data
through ad-hoc queries. In essence, managers
at every level can have a customized view
that extracts information from disparate
sources and summarizes it into meaningful
indicators. Executives need information for
strategic and tactical decision that often re-
quires the combination of data from ERP and
non-ERP application sources. The usual re-
ports developed from daily transactions does
not satisfy the business needs, an executive
cannot take a real time decision based on a
hundred pages per month cash-flow detailed
report. Information must be aggregated and
presented with a template based on a busi-
ness model. In the table below we represent
the main differences between ERP reports
and BIS reports [2]:
Table 1. A comparison between ERP and BIS reports.
Characteristics ERP reports BIS reports
Objectives Analyze indicators that meas-
ure current and internal activi-
ties or daily reports
Processes optimization, analyze key per-
formance indicators, forecast internal and
external data, internal and external focus
Level of decision Operational/Medium Strategic/High
User involved Operational level of manage-
ment
Executives, strategic level of management
Data Management Relational databases Data
warehouse
Data warehouse/OLAP/
Data Mining
Typical operation Report/Analyze Analyze
Number of records /
transaction
Limited Huge
Data Orientation Record Cube
Level of detail Detailed, summarized, pre-
aggregate
Aggregate
Age of data Current Historical/current/prospective
ERP systems are transaction-processing fo-
cused and weak on analytics. Strategic and
executive manager’s demand for technology
solutions that can extract, analyze, and vi-
sualize information from ERP and stand-
alone systems, and this has provided the mo-
tivation for a new type of information sys-
tems like BIS. The components of these sys-
tems are based on innovative technologies
such as data warehousing, OLAP, data min-
ing, friendly graphical user interfaces, inte-
grating tools capable of collecting,
processing, storing and retrieving data from
different sources. The following section de-
scribes some of these techniques and the ar-
chitecture of the BI systems.
3 Business Intelligence Systems’ Architec-
ture
BIS architecture is structured on three dis-
tinct levels [2]:
Level 1. Data Management – is represented
by relational databases, data warehouses and
other type of data sources. At this level is
common to use a data warehousing solution
that collects and organizes data from both in-
Informatica Economic? vol. 13, no. 4/2009 101
ternal and external sources and makes it
available for the purpose of analysis. A data
warehouse contains both historical and cur-
rent data and it is optimized for fast query
and analysis. Data warehouses extract, trans-
form and process data for high-level integra-
tion and analysis.
Although a data warehouse can make it easi-
er and more efficient to use the BIS, it is not
required for a BIS to be deployed. Organiza-
tions can extract data directly from their host
system database for their analysis and report-
ing purposes, but in a more difficult way.
Level 2. Model Management – is the level of
data extraction, transformation and
processing. This level is based on different
type of models for statistic interpretation,
analysis and forecasting data. At this level,
we can find technologies like OLAP, data
mining and analytical reporting. The OLAP
engine is a query generator that provides us-
ers with the ability to explore and analyze
summary and detailed information from a
multi-dimensional database. Traditional rela-
tional database systems handle this situation
by using multiple queries. In many cases, the
queries become so complex that even the de-
veloper finds them difficult to maintain.
OLAP overcomes this barrier by enabling
users to analyze multi-dimensional data.
Managers can use an OLAP engine or typical
operation like “slice and dice” data by vari-
ous dimensions and then drill down into the
source data or roll-up to aggregate levels.
OLAP provide tools for forecasting data and
“what-if” analysis. OLAP can only mark the
trends and patterns within the data that was
requested. It will not discover hidden rela-
tionships or patterns, which requires more
powerful tools like data mining (DM). These
tools are especially appropriate for large and
complex datasets. Through statistical or
modeling techniques, data mining tools make
it possible to discover hidden trends or rules
that are implicit in a large database. Data
mining tools can be applied to data from data
warehouses or relational databases. Data dis-
covered by these tools must be validated and
verified and then to become operational data
that can be used in decision process. OLAM
(on-line analytical data mining) systems are
OLAP systems used for data mining, used to
discover new information from multidimen-
sional-data.
Level 3. Data Visualization Tools - provide a
visual drill-down capacity that can help man-
agers examine data graphically and identify
complex interrelationships. BIS attempts to
present data in a form that is relevant for stra-
tegic decisions. At this level, one can find
tools for reporting and presenting data in a
friendly manner. A very efficient solution
that can be used also to integrate data is to
develop a business intelligence portal [4].
The main purpose of a BI portal is to inte-
grate data and information from a wide range
of applications and repositories, in order to
allow visualization of a multitude of systems,
either internal or external to organizations,
through a simple Web interface [5]. There-
fore, a BI portal can be seen like a Web-
based, secure interface, which can offer a
unique integration point for the applications
and services used by employees, partners,
suppliers and clients of the organization. The
main advantage of the information portal is
that it can be easily offered as a service to the
wide public [16].
4 Business Intelligence Systems’ Develop-
ment Lifecycle
There are some major differences between
transactional systems’ lifecycle and BIS life-
cycle which depends on decision systems’
characteristics, but the same traditional tech-
niques and stages are used for development:
initial study, project planning, analysis, de-
sign, construction, and implementation (fig-
ure 1).
In these stages there are many steps used for
modeling BIS characteristics such as [6]:
? orientation towards business opportuni-
ties rather than transactional needs;
? the implementation of strategically deci-
sions, not only departmental or opera-
tional decisions;
? analysis based on business needs, which
is the most important of the process;
? cyclical development process, focused on
evaluation and improvement of succes-
102 Informatica Economic? vol. 13, no. 4/2009
sive versions, not only building and ma- jor delivering of a singular a final version.
Fig. 1. BIS development lifecycle
BIS lifecycle is divided in 6 stages and 16
steps as following [3]:
Stage 1: J ustification
Step 1: Business case assessment - business
needs and opportunities are identified and
then the team proposes an initial solution jus-
tified by costs and benefits. A preliminary
report is built-up.
Stage 2: Planning
Step 2: Enterprise infrastructure evaluation –
this step estimates and values organization’s
capabilities to sustain and accomplish the
BIS project in terms of: infrastructure, com-
ponents, devices, network and also future
needs of these equipments. In this step is
built organization’s infrastructure.
Step 3: Project planning – BIS involves dy-
namical project planning which leads to rapid
changes in technology, organization and
business needs, human resources and imple-
menting team. The project plan is detailed,
progressive, each stage and step has checking
points and test documents and reports.
Stage 3: Business analysis
Step 4: Defining business needs – interviews
and meetings are organized with executives
and managers and business needs and re-
quirements are identified and defined. An ini-
tial solution is proposed, discussed and
adopted.
Step 5: Data analysis – this step involves
identifying and designing data sources, de-
signing detailed ER diagrams with attributes
and references between data. The logical
model is designed.
Step 6: Application prototyping – An initial
prototype is built and tested in order to vali-
date business needs. After testing results are
estimated and reported with positive and
negative aspects.
Step 7: Metadata analysis – metadata are de-
signed and data sources are mapped on meta-
data structure. CASE tools are used for de-
signing and mapping process.
Stage 4: System design
Step 8: Data design – in this step the logical
model is detailed and refined and physical
model is designed. The data model for
processing and storage are selected from the
following options: relational, object oriented
and multidimensional model.
Step 9: Designing the ETL process (extract /
transform / load) – this step is the most diffi-
cult in the entire cycle and depends on quali-
ty of data sources. It is recommended that the
process should be built in one environment
which integrate all modules of the organiza-
tion and not separately, on each department.
Informatica Economic? vol. 13, no. 4/2009 103
The rule should be: share one coordinated
ETL process.
Step 10: Design metadata repository – if it is
used a pre-defined solution for metadata re-
pository then in this step it is adjusted for
project requirements, otherwise a metadata
repository is designed in terms of metadata
logical model depending on data model: rela-
tional, object oriented or multidimensional.
Stage 5: Development
Step 11: ETL development – filtering tools,
procedures, operators are used for building
ETL process. Data filtering and transforma-
tions depends on data sources quality. These
sources are different like: files, databases, e-
mail, internet, unconventional sources.
Step 12: Application development – after
prototype validation, building the final appli-
cation may be a simple process. Procedures
templates and interfaces are re-built; user
rights and privileges are granted.
Step 13: Data Mining – executive systems
have to implement data mining capabilities in
order to succeed and accomplish manager’s
requirements. This step involves testing algo-
rithms, data mining techniques like clustering,
predictive and organizing methods.
Step 14: Developing metadata repository – if
the metadata repository has to be built-up
then metadata dictionary and data access in-
terfaces are developed.
Stage 6: System implementation
Step 15: Implementation – it is the delivering
process in which the development team or-
ganize training sessions for managers, final
documentations and technical support are
prepared, data loading process and applica-
tion setup is accomplished
Step 16: System testing – after system im-
plementation preliminary conclusions are
made, costs are estimated and the develop-
ment team build a final report in which are
describe system performances and also some
parts which have to be improved or re-built-
up.
5 Business Intelligence Conceptual Design
Model
In order to gather data from various sources
and ERP systems those are implemented in
an organization from different functional
areas or modules such as: financials, invento-
ry, purchase, order management, production
we need to analyze and design the business
model and strategic requests. This model
have to be mapped on a logical model and
physical model in the data warehouse and al-
so used for extracting and presenting data
through OLAP technology. These models are
known as multidimensional models and basi-
cally, they represent an extension of the rela-
tional model or ER schema or a multidimen-
sional view over facts.
Multidimensional models are classified in
two major types: models that are an exten-
sion of ER model are based on a star schema
and consist in the relationship between some
dimensions and facts or measures and n-
dimensional cube based models that use a
multidimensional view over an individual
situation or data.
In Business Intelligence Systems, the multi-
dimensional model that is used has to be able
to overhear the business requests. All we
need is a business vision over data structure
so the star schema or the n-cube based mod-
els have to design and incorporate business
aspects or demands not only the facts or the
relationship between data. The managers and
executives request a synthetic view over facts
and indicators and these key performance in-
dicators are built from the entire organiza-
tional data or even external data.
Also, the system have to provide a friendly
graphical interface with advanced capabili-
ties of slicing and dicing through data and
easily get a new perspective over data by ro-
tating dimensions and drill down or roll up
over hierarchical levels. So we need a multi-
dimensional model in which these operations
can be made easily, in real time and that can
it overhead the entire business model with re-
lationship between dimensions, facts and hie-
rarchies and it is based on the entire organi-
zational data at operational level, tactical lev-
el and strategically level.
Based on these considerations we propose an
extension of the star or the constellation
schema but with aggregate data and hierar-
chies in fact tables not only in dimension
104 Informatica Economic? vol. 13, no. 4/2009
tables. The model is structured over three
distinct levels and we can call it a pyramidal
model with the following structure (for de-
tailed description see the book [2]):
? Organizational level (or the base of the
pyramid) – containing dimensions and
facts with an organizational scope, at a
general level, that shape and are common
to the entire activities. Such dimensions
can be: , , ,
and facts: production, pur-
chasing etc. Data are at a detailed level
with multiple hierarchies over each di-
mension table.
? Departmental level – containing dimen-
sions and facts for the departmental le-
vels of the organization and particular ac-
tivities in these departments or field of in-
terests, group by data marts or data cen-
ters. Such dimensions can be: ,
, and facts: stocks,
payments, sales etc. Data are at a detailed
and aggregate level with specialized hie-
rarchies over each dimension table.
? Strategically level – containing dimen-
sions and facts derived from the base di-
mensions and facts, with specific ele-
ments for the strategic analysis, like , , and
facts: cash-flow, KPIs. Data are at an ag-
gregate, synthetic level with specialized
hierarchies over each dimension table.
The main characteristic of the model is that
between the dimension tables and the facts
from different levels of the architecture can
be establish a relationship and also the fact
tables can have hierarchies and class
attributes that can be used for drill down or
roll up.
Advantages of the model:
Flexibility – new elements or objects like
new dimensions or facts can easily be in-
cluded in the model without affecting the ex-
isting architecture or remodeling the system
and the loading process for a specific level
can be made without refreshing the whole da-
ta;
Real model of business requirements – the
three level architecture is based on the real
model of business requirements thus this
model can be mapped on the each level of the
pyramid;
Performance in the drill-down or roll-up op-
erations – because the dimensions and facts
are separated at each level we can easily na-
vigate through hierarchies from a level to
another;
Incremental development – the model can be
built in stages and each stage can be vali-
dated and used before the next stage;
EIS, MIS and DSS support – the top level can
be used to implement an Executive Informa-
tion System (EIS), the bottom and middle le-
vels can be used for design and realized a
Management Information System (MIS) or a
Decision Support System (DSS) because
these systems can use the specific dimension
and fact tables from these levels.
Disadvantages of the model:
High complexity – because it is containing
three different levels, the business model
need to be careful analyzed and designed in
order to identify the proper and suitable di-
mensions and facts and also the hierarchies at
each level. An inadequate choice can have a
major effect on the performance of the entire
system;
Moderate performance of the interrogation
process – in order to perform a complex
query the model need to establish many rela-
tionships and joins between the fact and di-
mension tables and this can reduce the per-
formance of interrogation;
Top-down and bottom-up development – In
order to overhear the entire aspects of the
business process we need to build the sys-
tems in two directions: first top-bottom to
model the strategic requirements and second,
bottom-up for validating and setting up the
hierarchical flux of data.
The pyramidal model is suited for business
needs and can be developed and imple-
mented through an object oriented approach,
defining classes for dimensions and facts,
following the rules of the OO design. The
prototype will have the main functionalities
of the business model and when there will be
any change in this model and a new business
requirement appear then new functions or
new attributes can be added to the main
Informatica Economic? vol. 13, no. 4/2009 105
classes to complete the demands.
6 Criteria for Evaluating Business Intelli-
gence Systems
A problem of BI systems is measuring suc-
cess. There are some case studies evidencing
benefits generated by organizations that are
successful with the use of BIS like the one
shown in [8] and [9], but with limited empir-
ically validated measures. There have been
several studies calling for the development
of a measure for evaluating the business per-
formance effects of BIS ([10], [11], [12]).
In deploying a BIS there are many risks in-
volved: system design, data quality, and
technology obsolescence. System design
risks stem from poor conceptualization of an
enterprise’s true business needs before the
technology is deployed. Data quality risks re-
late primarily to whether or not data has been
properly cleansed. Technology obsolescence
refers to the failure on the part of the vendor
to anticipate new technologies.
Large budgets and strategic information are
involved in deploying BIS systems – this is
the reason to establish rigorous criteria for
evaluating such systems. These criteria are
discussed below.
Decisions based on business process
BIS should not be viewed only as a data re-
pository or a large set of data. Instead, sys-
tem’s implementation should be concern on
conceptualizing new data models, processes,
and indicators that form the content of BIS;
also it should provide extensive understand-
ing of the benchmarks that are useful to eva-
luate business processes.
Performance
This feature typically refers to the response
time that a system provides to its users. Most
responses should range from a few seconds
to a maximum of 30 seconds for routine que-
ries. Response times depend on the com-
plexity of the database and the queries being
requested.
Flexibility and scalability
Flexibility determines whether a BI solution
can continually adapt to changing business
conditions after the system has been deli-
vered. BIS should be able to accommodate
changes in any type of business process and
functions like personnel, services, and
processes, as well as new mandates, laws,
and regulations requiring the capture of dif-
ferent types of data. BIS should be expanda-
ble to accommodate data growth and changes
to organizational structure. EIS also should
allow contributed content to grow without a
slowdown in performance.
Integration
Integration involves two types of issues: data
integration and system integration.
Data integration is the ability to access data
from much different type of systems, so BIS
will be particularly effective if it can over-
come the challenge of information fragmen-
tation, allowing executives to measure fea-
tures of business processes that involve in-
formation from inside and outside of the or-
ganization. System integration refers to two
things: the ability to extent the BI software
with new capabilities and modules and the
system’s ability to coexist with other enter-
prise solutions.
Friendly user interface
BIS should be designed to allow managers
who are not trained to use query languages
and advanced technologies, a fast, easy, and
understandable way to navigate into data and
identify trends and patterns. BIS should per-
mit the user interface to accommodate differ-
ent degrees of technical knowledge.
7 Conclusions
BI systems have a powerful impact on stra-
tegic decisions quality to reduce the time for
making decisions and thus these systems
must have the ability to allow managers to
view data in different perspective, to drill-
down and roll-up to aggregate levels, to na-
vigate and on-line query data sets in order to
discover new factors that affect business
process and also to anticipate and forecast
changes inside and outside the organization.
BIS improve the quality of management in
organization through new type of technology
and techniques for extracting, transforming,
processing and presenting data in order to
provide strategic information.
One of the major risks in the process of de-
106 Informatica Economic? vol. 13, no. 4/2009
veloping a BIS is the system design that
stem from poor conceptualization of an en-
terprise’s true business needs before the sys-
tems is deployed and for every change in
these requirements the prototype must be al-
so revised. A solution for covering this risk
is object oriented modeling of a data ware-
house that helps us to improve the designing
phase and the development cycle and also
we can re-use some parts of the prototype
that it was implemented in an organization
in order to design and implement another
prototype in other organizational environ-
ments. So, object types can store structured
business data in its natural form in object
tables and then allow applications, such as
OLAP applications, to work in a multidi-
mensional way using the object oriented
properties and facilities.
Acknowledgement
This paper presents some results of the re-
search project PN II, Ideas Program, Code
820: Informatics solutions for decision sup-
port and for the development of knowledge
management in public institutions, financed
within the framework of IDEI research pro-
gram.
This article is a result of the project Doctoral
Program and PhD Students in the education
research and innovation triangle. This
project is co funded by European Social Fund
through The Sectoral Operational Program
for Human Resources Development 2007-
2013, coordinated by The Bucharest Acade-
my of Economic Studies.
References
[1] S. Luján-Mora and J . Trujillo, “Extend-
ing UML for Multidimensional Model-
ling,” Proceedings of the 5th Interna-
tional Conference on the Unified Model-
ling Language, Germany, 2002.
[2] I. Lungu and A. Bara, Executive Infor-
mation Systems, ASE Printing House,
Bucharest, 2007.
[3] L. Moss and S. Atre, Business Intelli-
gence Roadmap – The complete project
lifecycle for decision-support applica-
tions, Addison-Wesley, 2004.
[4] A. R. Bologa, R. Bologa and A. Bara,
“Technology vs Business Needs in Busi-
ness Intelligence Projects,” Proceedings
of the International Conference on e-
Business (ICE-B 2008), 26-29 J uly, 2008,
Porto, Portugal.
[5] V. Diaconita, I. Botha, A. Bâra, I. Lungu
and M. Velicanu, “Two Integration Fla-
vors in Public Institutions,” WSEAS
Transactions on Information Science and
Applications, May 2008.
[6] L. Y. Liang and R. Miranda, “Dash-
boards and Scorecards: Executive Infor-
mation Systems for the Public Sector,”
Government Finance Review, December
2001.
[7] M. Z. Elbashir, P. A. Collier and M. J .
Davern, “Measuring the effects of busi-
ness intelligence systems - The relation-
ship between business process and orga-
nizational performance,” International
Journal of Accounting Information Sys-
tems, pp. 135–153, 2008.
[8] J . Hesford and K. Antia, A process-
oriented view of competitive intelligence
and its impact on organizational perfor-
mance, Working Paper. Richard Ivey
School of Business, The University of
Western Ontario, 2006.
[9] M. Edwards, “Best practices and leader-
ship in data warehousing award winners
2002,” Bus Intell J, Vol. 7, No. 4, 2002.
[10] A. Lönnqvist and V. Pirttimäki, “The
measurement of business intelligence,”
Information Systems Management, Vol.
23, No. 1, pp. 32–40, 2006.
[11] J . Marin and A. Poulter, “Dissemina-
tion of competitive intelligence,” Journal
of Information Science, Vol. 30, No. 2,
pp. 165–80, 2004.
[12] L. Davison, “Measuring competitive in-
telligence effectiveness: Insights from
the advertising industry,” Competitive
Intelligence Review, Vol. 12, No. 4, 2001,
pp. 25–38.
[13] Y. C. Lu and H. Cheng, Automated op-
timal equity portfolios discovery in a fi-
nancial knowledge management system,
computational intelligence in economics
and finance, Springer-Verlag, 2003, pp.
Informatica Economic? vol. 13, no. 4/2009 107
387–402.
[14] Z. Wang, Business intelligence, DrMas-
ter Culture Limited Company, 2005.
[15] T. Davenport, “Competing on Analyt-
ics,” Harvard Business Review, 2005.
[16] I. Lungu, M. Velicanu, A. Bara, V. Di-
aconita and I. Botha, “Portal based sys-
tem integration – foundation for decision
support,” Economic Computation and
Economic Cybernetics Studies and Re-
search Journal, Vol. 43, No. 1, 2009, pp.
123-135.
Adela BÂRA is a Lecturer at the Economic Informatics Department at the
Faculty of Cybernetics, Statistics and Economic Informatics from the Acad-
emy of Economic Studies of Bucharest. She has graduated the Faculty of
Economic Cybernetics in 2002, holds a PhD diploma in Economics from
2007. She is the author of 7 books in the domain of economic informatics,
over 40 published scientific papers and articles (among which over 20 articles
are indexed in international databases, ISI proceedings, SCOPUS and 2 of
them are ISI indexed). She participated as team member in 3 research projects that have been
financed from national research programs. She is a member of INFOREC professional associ-
ation. From May 2009, she is the director of the Oracle Excellence Centre in the university,
responsible for the implementation of the Oracle Academy Initiative program. Domains of
competence: Database systems, Data warehouses, OLAP and Business Intelligence, Executive
Information Systems, Decision Support Systems, Data Mining.
Iuliana BOTHA is an Assistant Lecturer at the Economic Informatics De-
partment at the Faculty of Cybernetics, Statistics and Economic Informatics
from the Academy of Economic Studies of Bucharest. She has graduated the
Faculty of Cybernetics, Statistics and Economic Informatics in 2006 and the
Databases for Business Support master program organized by the Academy
of Economic Studies of Bucharest in 2008. Currently, she is a PhD student in
the field of Economic Informatics at the Academy of Economic Studies. She
is co-author of two books, 13 published scientific papers and articles (among which two pa-
pers are ISI indexed and another 5 are included in international databases). She participated as
team member in 3 research projects that have been financed from national research programs.
From 2007, she is the scientific secretary of the master program Databases for Business Sup-
port and she is also a member of INFOREC professional association. Her scientific fields of
interest include: Database Systems, Design of Economic Information Systems, Grid Compu-
ting, e-Learning Technologies.
Vlad DIACONI?A is an Assistant Lecturer at the Economic Informatics
Department at the Faculty of Cybernetics, Statistics and Economic Informat-
ics from the Academy of Economic Studies of Bucharest. He has graduated
the faculty at which he is now teaching in 2005 and studies for his PhD in the
field of Cybernetics and Statistics. He is the co-author of 2 books in the do-
main of economic informatics, 2 articles in ISI journals, 4 articles in Scopus
journals, 4 articles in ISI proceedings, 4 papers in B+journals and 6 papers in
the proceedings of international conferences. He participated as team member in 3 research
projects that have been financed from national research programs. He is a member of INFO-
REC professional association. Domains of competence: Database systems, Data warehouses,
OLAP and Business Intelligence, Integrated Systems, SOA.
108 Informatica Economic? vol. 13, no. 4/2009
Ion LUNGU is a Professor at the Economic Informatics Department at the
Faculty of Cybernetics, Statistics and Economic Informatics from the Acad-
emy of Economic Studies of Bucharest. He has graduated the Faculty of Eco-
nomic Cybernetics in 1974, holds a PhD diploma in Economics from 1983
and, starting with 1999 is a PhD coordinator in the field of Economic Infor-
matics. He is the author of 22 books in the domain of economic informatics,
57 published articles (among which 2 articles ISI indexed) and 39 scientific
papers published in conferences proceedings (among which 5 papers ISI indexed and 15 in-
cluded in international databases). He participated (as director or as team member) in more
than 20 research projects that have been financed from national research programs. He is a
CNCSIS expert evaluator and member of the scientific board for the ISI indexed journal Eco-
nomic Computation and Economic Cybernetics Studies and Research. He is also a member of
INFOREC professional association and honorific member of Economic Independence aca-
demic association. In 2005 he founded the master program Databases for Business Support
(classic and online), who’s manager he is. His fields of interest include: Databases, Design of
Economic Information Systems, Database Management Systems, Decision Support Systems,
Executive Information Systems.
Anda VELICANU has graduated the Faculty of Economic Cybernetics, Sta-
tistics and Informatics of the Bucharest Academy of Economic Studies, in
2008. She is a PhD student in the field of Economic Informatics at the Acad-
emy of Economic Studies and since J anuary 2009, she is a Pre-Assistant
Lecturer. She teaches Database, Database Management Systems and Eco-
nomic Informatics seminars at the following faculties: Economic Cybernetics,
Statistics and Informatics, Commerce, Marketing and International Business
and Economics. Her research activity can be observed in the following achievements: 5 dip-
lomas, 2 scientific awards, 3 proceedings, 2 articles published in scientific reviews, 1 research
contract, 1 book and 1 research grant. She is a member of INFOREC professional association.
Her scientific fields of interest include: Databases, Database Management Systems, Pro-
gramming, Information Systems.
Manole VELICANU is a Professor at the Economic Informatics Department
at the Faculty of Cybernetics, Statistics and Economic Informatics from the
Academy of Economic Studies of Bucharest. He has graduated the Faculty of
Economic Cybernetics in 1976, holds a PhD diploma in Economics from
1994 and starting with 2002 he is a PhD coordinator in the field of Economic
Informatics. He is the author of 18 books in the domain of economic infor-
matics, 64 published articles (among which 2 articles ISI indexed), 55 scien-
tific papers published in conferences proceedings (among which 5 papers ISI indexed and 7
included in international databases) and 36 scientific papers presented at conferences, but un-
published. He participated (as director or as team member) in more than 40 research projects
that have been financed from national research programs. He is a member of INFOREC pro-
fessional association, a CNCSIS expert evaluator and a MCT expert evaluator for the program
Cercetare de Excelenta - CEEX (from 2006). From 2005 he is co-manager of the master pro-
gram Databases for Business Support. His fields of interest include: Databases, Design of
Economic Information Systems, Database Management Systems, Artificial Intelligence, Pro-
gramming languages.
doc_385719315.pdf
Business Intelligence Systems (BIS) require historical data or data collected from various
sources.
Informatica Economic? vol. 13, no. 4/2009 99
A model for Business Intelligence Systems’ Development
Adela BARA, Iuliana BOTHA, Vlad DIACONI?A, Ion LUNGU, Anda VELICANU,
Manole VELICANU
Academy of Economic Studies, Bucharest, Romania
Faculty of Economics Cybernetics, Statistics and Informatics
[email protected], [email protected], [email protected],
[email protected], [email protected], [email protected]
Often, Business Intelligence Systems (BIS) require historical data or data collected from var-
ious sources. The solution is found in data warehouses, which are the main technology used
to extract, transform, load and store data in the organizational Business Intelligence projects.
The development cycle of a data warehouse involves lots of resources, time, high costs and
above all, it is built only for some specific tasks. In this paper, we’ll present some of the as-
pects of the BI systems’ development such as: architecture, lifecycle, modeling techniques and
finally, some evaluation criteria for the system’s performance.
Keywords: BIS (Business Intelligence Systems), Data Warehouses, OLAP (On-Line Analytical
Processing), Object-Oriented Modeling
Introduction
The main objective of our research
project that we had in one of the multination-
al companies from Romania is to develop a
decision support system for public institution,
so we tried to meet the requests from the ex-
ecutives and managers of one national com-
pany. We based our work on previous expe-
riences, researches, articles and studies that
we’d been developed. Thus, we followed the
next classical steps: analyze, design, develop
and applying for the project lifecycle the
framework described in the book [2]. For the
implementation phase we used different BI
techniques, like data warehousing, OLAP,
data mining, portal and we finally succeeded
to implement the BI system’s prototype and
to validate it with the managers and execu-
tives in one national company. The system
gathers data, using the ERP system, to extract
data from different functional areas or mod-
ules such as: financials, inventory, purchase,
order management or production. For the ex-
ecutives, the system is able to provide analyt-
ical reports and dashboards. As the storage
solution we designed and build a data ware-
house. The major problem is that there will
be many more changes in the structure of the
organization and the impact of these changes
may affect the BI system. So, we need to find
a solution, and, based on our previous re-
searches on Object Oriented modeling, we
consider it as a good option. In the last years,
there have been some proposals to represent
MD properties at the conceptual level. As we
mentioned in our previous researches, based
on the conference paper [1], we defined a set
of object-oriented extensions that can be used
for modeling the components and require-
ments of a data warehouse. Also, we had to
consider the system’s development lifecycle
that has to be flexible and easy to fulfill.
2 The concept of Business Intelligence Sys-
tem
With rapid advances in technology, enter-
prises today frequently search for new ways
to establish value positions. Well built Busi-
ness Intelligence Systems (BIS) can provide
the ability to analyze business information in
order to support and improve management
decision making across a broad range of
business activities. They leverage the large
data infrastructure investment, for example,
in ERP systems made by firms, and have the
potential to realize the substantial value
locked up in a firm’s data resources [7].
While substantial business investment in BI
systems is continuing to accelerate, there is a
complete absence of a specific and rigorous
method to measure the realized business val-
ue, if any. BI systems have the potential to
1
100 Informatica Economic? vol. 13, no. 4/2009
maximize the use of information by improv-
ing the company’s capacity to structure a
large volume of information and make it ac-
cessible, thereby creating competitive advan-
tage: “competing on analytics” [15].
BI utilizes a substantial amount of collected
data during the daily operational processes,
and transforms the data into information and
knowledge to avoid the supposition and ig-
norance of the enterprises [14].
The main characteristics of a BIS are: the ca-
pability of providing representative informa-
tion to the high-level management, to support
strategic activities such as goal setting, plan-
ning and forecasting, and also tracking per-
formance, to gather, analyze, and integrate
internal and external data into dynamic pro-
files of key performance indicators. Based on
each executive’s information needs, BIS can
access both historical and real-time data
through ad-hoc queries. In essence, managers
at every level can have a customized view
that extracts information from disparate
sources and summarizes it into meaningful
indicators. Executives need information for
strategic and tactical decision that often re-
quires the combination of data from ERP and
non-ERP application sources. The usual re-
ports developed from daily transactions does
not satisfy the business needs, an executive
cannot take a real time decision based on a
hundred pages per month cash-flow detailed
report. Information must be aggregated and
presented with a template based on a busi-
ness model. In the table below we represent
the main differences between ERP reports
and BIS reports [2]:
Table 1. A comparison between ERP and BIS reports.
Characteristics ERP reports BIS reports
Objectives Analyze indicators that meas-
ure current and internal activi-
ties or daily reports
Processes optimization, analyze key per-
formance indicators, forecast internal and
external data, internal and external focus
Level of decision Operational/Medium Strategic/High
User involved Operational level of manage-
ment
Executives, strategic level of management
Data Management Relational databases Data
warehouse
Data warehouse/OLAP/
Data Mining
Typical operation Report/Analyze Analyze
Number of records /
transaction
Limited Huge
Data Orientation Record Cube
Level of detail Detailed, summarized, pre-
aggregate
Aggregate
Age of data Current Historical/current/prospective
ERP systems are transaction-processing fo-
cused and weak on analytics. Strategic and
executive manager’s demand for technology
solutions that can extract, analyze, and vi-
sualize information from ERP and stand-
alone systems, and this has provided the mo-
tivation for a new type of information sys-
tems like BIS. The components of these sys-
tems are based on innovative technologies
such as data warehousing, OLAP, data min-
ing, friendly graphical user interfaces, inte-
grating tools capable of collecting,
processing, storing and retrieving data from
different sources. The following section de-
scribes some of these techniques and the ar-
chitecture of the BI systems.
3 Business Intelligence Systems’ Architec-
ture
BIS architecture is structured on three dis-
tinct levels [2]:
Level 1. Data Management – is represented
by relational databases, data warehouses and
other type of data sources. At this level is
common to use a data warehousing solution
that collects and organizes data from both in-
Informatica Economic? vol. 13, no. 4/2009 101
ternal and external sources and makes it
available for the purpose of analysis. A data
warehouse contains both historical and cur-
rent data and it is optimized for fast query
and analysis. Data warehouses extract, trans-
form and process data for high-level integra-
tion and analysis.
Although a data warehouse can make it easi-
er and more efficient to use the BIS, it is not
required for a BIS to be deployed. Organiza-
tions can extract data directly from their host
system database for their analysis and report-
ing purposes, but in a more difficult way.
Level 2. Model Management – is the level of
data extraction, transformation and
processing. This level is based on different
type of models for statistic interpretation,
analysis and forecasting data. At this level,
we can find technologies like OLAP, data
mining and analytical reporting. The OLAP
engine is a query generator that provides us-
ers with the ability to explore and analyze
summary and detailed information from a
multi-dimensional database. Traditional rela-
tional database systems handle this situation
by using multiple queries. In many cases, the
queries become so complex that even the de-
veloper finds them difficult to maintain.
OLAP overcomes this barrier by enabling
users to analyze multi-dimensional data.
Managers can use an OLAP engine or typical
operation like “slice and dice” data by vari-
ous dimensions and then drill down into the
source data or roll-up to aggregate levels.
OLAP provide tools for forecasting data and
“what-if” analysis. OLAP can only mark the
trends and patterns within the data that was
requested. It will not discover hidden rela-
tionships or patterns, which requires more
powerful tools like data mining (DM). These
tools are especially appropriate for large and
complex datasets. Through statistical or
modeling techniques, data mining tools make
it possible to discover hidden trends or rules
that are implicit in a large database. Data
mining tools can be applied to data from data
warehouses or relational databases. Data dis-
covered by these tools must be validated and
verified and then to become operational data
that can be used in decision process. OLAM
(on-line analytical data mining) systems are
OLAP systems used for data mining, used to
discover new information from multidimen-
sional-data.
Level 3. Data Visualization Tools - provide a
visual drill-down capacity that can help man-
agers examine data graphically and identify
complex interrelationships. BIS attempts to
present data in a form that is relevant for stra-
tegic decisions. At this level, one can find
tools for reporting and presenting data in a
friendly manner. A very efficient solution
that can be used also to integrate data is to
develop a business intelligence portal [4].
The main purpose of a BI portal is to inte-
grate data and information from a wide range
of applications and repositories, in order to
allow visualization of a multitude of systems,
either internal or external to organizations,
through a simple Web interface [5]. There-
fore, a BI portal can be seen like a Web-
based, secure interface, which can offer a
unique integration point for the applications
and services used by employees, partners,
suppliers and clients of the organization. The
main advantage of the information portal is
that it can be easily offered as a service to the
wide public [16].
4 Business Intelligence Systems’ Develop-
ment Lifecycle
There are some major differences between
transactional systems’ lifecycle and BIS life-
cycle which depends on decision systems’
characteristics, but the same traditional tech-
niques and stages are used for development:
initial study, project planning, analysis, de-
sign, construction, and implementation (fig-
ure 1).
In these stages there are many steps used for
modeling BIS characteristics such as [6]:
? orientation towards business opportuni-
ties rather than transactional needs;
? the implementation of strategically deci-
sions, not only departmental or opera-
tional decisions;
? analysis based on business needs, which
is the most important of the process;
? cyclical development process, focused on
evaluation and improvement of succes-
102 Informatica Economic? vol. 13, no. 4/2009
sive versions, not only building and ma- jor delivering of a singular a final version.
Fig. 1. BIS development lifecycle
BIS lifecycle is divided in 6 stages and 16
steps as following [3]:
Stage 1: J ustification
Step 1: Business case assessment - business
needs and opportunities are identified and
then the team proposes an initial solution jus-
tified by costs and benefits. A preliminary
report is built-up.
Stage 2: Planning
Step 2: Enterprise infrastructure evaluation –
this step estimates and values organization’s
capabilities to sustain and accomplish the
BIS project in terms of: infrastructure, com-
ponents, devices, network and also future
needs of these equipments. In this step is
built organization’s infrastructure.
Step 3: Project planning – BIS involves dy-
namical project planning which leads to rapid
changes in technology, organization and
business needs, human resources and imple-
menting team. The project plan is detailed,
progressive, each stage and step has checking
points and test documents and reports.
Stage 3: Business analysis
Step 4: Defining business needs – interviews
and meetings are organized with executives
and managers and business needs and re-
quirements are identified and defined. An ini-
tial solution is proposed, discussed and
adopted.
Step 5: Data analysis – this step involves
identifying and designing data sources, de-
signing detailed ER diagrams with attributes
and references between data. The logical
model is designed.
Step 6: Application prototyping – An initial
prototype is built and tested in order to vali-
date business needs. After testing results are
estimated and reported with positive and
negative aspects.
Step 7: Metadata analysis – metadata are de-
signed and data sources are mapped on meta-
data structure. CASE tools are used for de-
signing and mapping process.
Stage 4: System design
Step 8: Data design – in this step the logical
model is detailed and refined and physical
model is designed. The data model for
processing and storage are selected from the
following options: relational, object oriented
and multidimensional model.
Step 9: Designing the ETL process (extract /
transform / load) – this step is the most diffi-
cult in the entire cycle and depends on quali-
ty of data sources. It is recommended that the
process should be built in one environment
which integrate all modules of the organiza-
tion and not separately, on each department.
Informatica Economic? vol. 13, no. 4/2009 103
The rule should be: share one coordinated
ETL process.
Step 10: Design metadata repository – if it is
used a pre-defined solution for metadata re-
pository then in this step it is adjusted for
project requirements, otherwise a metadata
repository is designed in terms of metadata
logical model depending on data model: rela-
tional, object oriented or multidimensional.
Stage 5: Development
Step 11: ETL development – filtering tools,
procedures, operators are used for building
ETL process. Data filtering and transforma-
tions depends on data sources quality. These
sources are different like: files, databases, e-
mail, internet, unconventional sources.
Step 12: Application development – after
prototype validation, building the final appli-
cation may be a simple process. Procedures
templates and interfaces are re-built; user
rights and privileges are granted.
Step 13: Data Mining – executive systems
have to implement data mining capabilities in
order to succeed and accomplish manager’s
requirements. This step involves testing algo-
rithms, data mining techniques like clustering,
predictive and organizing methods.
Step 14: Developing metadata repository – if
the metadata repository has to be built-up
then metadata dictionary and data access in-
terfaces are developed.
Stage 6: System implementation
Step 15: Implementation – it is the delivering
process in which the development team or-
ganize training sessions for managers, final
documentations and technical support are
prepared, data loading process and applica-
tion setup is accomplished
Step 16: System testing – after system im-
plementation preliminary conclusions are
made, costs are estimated and the develop-
ment team build a final report in which are
describe system performances and also some
parts which have to be improved or re-built-
up.
5 Business Intelligence Conceptual Design
Model
In order to gather data from various sources
and ERP systems those are implemented in
an organization from different functional
areas or modules such as: financials, invento-
ry, purchase, order management, production
we need to analyze and design the business
model and strategic requests. This model
have to be mapped on a logical model and
physical model in the data warehouse and al-
so used for extracting and presenting data
through OLAP technology. These models are
known as multidimensional models and basi-
cally, they represent an extension of the rela-
tional model or ER schema or a multidimen-
sional view over facts.
Multidimensional models are classified in
two major types: models that are an exten-
sion of ER model are based on a star schema
and consist in the relationship between some
dimensions and facts or measures and n-
dimensional cube based models that use a
multidimensional view over an individual
situation or data.
In Business Intelligence Systems, the multi-
dimensional model that is used has to be able
to overhear the business requests. All we
need is a business vision over data structure
so the star schema or the n-cube based mod-
els have to design and incorporate business
aspects or demands not only the facts or the
relationship between data. The managers and
executives request a synthetic view over facts
and indicators and these key performance in-
dicators are built from the entire organiza-
tional data or even external data.
Also, the system have to provide a friendly
graphical interface with advanced capabili-
ties of slicing and dicing through data and
easily get a new perspective over data by ro-
tating dimensions and drill down or roll up
over hierarchical levels. So we need a multi-
dimensional model in which these operations
can be made easily, in real time and that can
it overhead the entire business model with re-
lationship between dimensions, facts and hie-
rarchies and it is based on the entire organi-
zational data at operational level, tactical lev-
el and strategically level.
Based on these considerations we propose an
extension of the star or the constellation
schema but with aggregate data and hierar-
chies in fact tables not only in dimension
104 Informatica Economic? vol. 13, no. 4/2009
tables. The model is structured over three
distinct levels and we can call it a pyramidal
model with the following structure (for de-
tailed description see the book [2]):
? Organizational level (or the base of the
pyramid) – containing dimensions and
facts with an organizational scope, at a
general level, that shape and are common
to the entire activities. Such dimensions
can be: , , ,
and facts: production, pur-
chasing etc. Data are at a detailed level
with multiple hierarchies over each di-
mension table.
? Departmental level – containing dimen-
sions and facts for the departmental le-
vels of the organization and particular ac-
tivities in these departments or field of in-
terests, group by data marts or data cen-
ters. Such dimensions can be: ,
, and facts: stocks,
payments, sales etc. Data are at a detailed
and aggregate level with specialized hie-
rarchies over each dimension table.
? Strategically level – containing dimen-
sions and facts derived from the base di-
mensions and facts, with specific ele-
ments for the strategic analysis, like , , and
facts: cash-flow, KPIs. Data are at an ag-
gregate, synthetic level with specialized
hierarchies over each dimension table.
The main characteristic of the model is that
between the dimension tables and the facts
from different levels of the architecture can
be establish a relationship and also the fact
tables can have hierarchies and class
attributes that can be used for drill down or
roll up.
Advantages of the model:
Flexibility – new elements or objects like
new dimensions or facts can easily be in-
cluded in the model without affecting the ex-
isting architecture or remodeling the system
and the loading process for a specific level
can be made without refreshing the whole da-
ta;
Real model of business requirements – the
three level architecture is based on the real
model of business requirements thus this
model can be mapped on the each level of the
pyramid;
Performance in the drill-down or roll-up op-
erations – because the dimensions and facts
are separated at each level we can easily na-
vigate through hierarchies from a level to
another;
Incremental development – the model can be
built in stages and each stage can be vali-
dated and used before the next stage;
EIS, MIS and DSS support – the top level can
be used to implement an Executive Informa-
tion System (EIS), the bottom and middle le-
vels can be used for design and realized a
Management Information System (MIS) or a
Decision Support System (DSS) because
these systems can use the specific dimension
and fact tables from these levels.
Disadvantages of the model:
High complexity – because it is containing
three different levels, the business model
need to be careful analyzed and designed in
order to identify the proper and suitable di-
mensions and facts and also the hierarchies at
each level. An inadequate choice can have a
major effect on the performance of the entire
system;
Moderate performance of the interrogation
process – in order to perform a complex
query the model need to establish many rela-
tionships and joins between the fact and di-
mension tables and this can reduce the per-
formance of interrogation;
Top-down and bottom-up development – In
order to overhear the entire aspects of the
business process we need to build the sys-
tems in two directions: first top-bottom to
model the strategic requirements and second,
bottom-up for validating and setting up the
hierarchical flux of data.
The pyramidal model is suited for business
needs and can be developed and imple-
mented through an object oriented approach,
defining classes for dimensions and facts,
following the rules of the OO design. The
prototype will have the main functionalities
of the business model and when there will be
any change in this model and a new business
requirement appear then new functions or
new attributes can be added to the main
Informatica Economic? vol. 13, no. 4/2009 105
classes to complete the demands.
6 Criteria for Evaluating Business Intelli-
gence Systems
A problem of BI systems is measuring suc-
cess. There are some case studies evidencing
benefits generated by organizations that are
successful with the use of BIS like the one
shown in [8] and [9], but with limited empir-
ically validated measures. There have been
several studies calling for the development
of a measure for evaluating the business per-
formance effects of BIS ([10], [11], [12]).
In deploying a BIS there are many risks in-
volved: system design, data quality, and
technology obsolescence. System design
risks stem from poor conceptualization of an
enterprise’s true business needs before the
technology is deployed. Data quality risks re-
late primarily to whether or not data has been
properly cleansed. Technology obsolescence
refers to the failure on the part of the vendor
to anticipate new technologies.
Large budgets and strategic information are
involved in deploying BIS systems – this is
the reason to establish rigorous criteria for
evaluating such systems. These criteria are
discussed below.
Decisions based on business process
BIS should not be viewed only as a data re-
pository or a large set of data. Instead, sys-
tem’s implementation should be concern on
conceptualizing new data models, processes,
and indicators that form the content of BIS;
also it should provide extensive understand-
ing of the benchmarks that are useful to eva-
luate business processes.
Performance
This feature typically refers to the response
time that a system provides to its users. Most
responses should range from a few seconds
to a maximum of 30 seconds for routine que-
ries. Response times depend on the com-
plexity of the database and the queries being
requested.
Flexibility and scalability
Flexibility determines whether a BI solution
can continually adapt to changing business
conditions after the system has been deli-
vered. BIS should be able to accommodate
changes in any type of business process and
functions like personnel, services, and
processes, as well as new mandates, laws,
and regulations requiring the capture of dif-
ferent types of data. BIS should be expanda-
ble to accommodate data growth and changes
to organizational structure. EIS also should
allow contributed content to grow without a
slowdown in performance.
Integration
Integration involves two types of issues: data
integration and system integration.
Data integration is the ability to access data
from much different type of systems, so BIS
will be particularly effective if it can over-
come the challenge of information fragmen-
tation, allowing executives to measure fea-
tures of business processes that involve in-
formation from inside and outside of the or-
ganization. System integration refers to two
things: the ability to extent the BI software
with new capabilities and modules and the
system’s ability to coexist with other enter-
prise solutions.
Friendly user interface
BIS should be designed to allow managers
who are not trained to use query languages
and advanced technologies, a fast, easy, and
understandable way to navigate into data and
identify trends and patterns. BIS should per-
mit the user interface to accommodate differ-
ent degrees of technical knowledge.
7 Conclusions
BI systems have a powerful impact on stra-
tegic decisions quality to reduce the time for
making decisions and thus these systems
must have the ability to allow managers to
view data in different perspective, to drill-
down and roll-up to aggregate levels, to na-
vigate and on-line query data sets in order to
discover new factors that affect business
process and also to anticipate and forecast
changes inside and outside the organization.
BIS improve the quality of management in
organization through new type of technology
and techniques for extracting, transforming,
processing and presenting data in order to
provide strategic information.
One of the major risks in the process of de-
106 Informatica Economic? vol. 13, no. 4/2009
veloping a BIS is the system design that
stem from poor conceptualization of an en-
terprise’s true business needs before the sys-
tems is deployed and for every change in
these requirements the prototype must be al-
so revised. A solution for covering this risk
is object oriented modeling of a data ware-
house that helps us to improve the designing
phase and the development cycle and also
we can re-use some parts of the prototype
that it was implemented in an organization
in order to design and implement another
prototype in other organizational environ-
ments. So, object types can store structured
business data in its natural form in object
tables and then allow applications, such as
OLAP applications, to work in a multidi-
mensional way using the object oriented
properties and facilities.
Acknowledgement
This paper presents some results of the re-
search project PN II, Ideas Program, Code
820: Informatics solutions for decision sup-
port and for the development of knowledge
management in public institutions, financed
within the framework of IDEI research pro-
gram.
This article is a result of the project Doctoral
Program and PhD Students in the education
research and innovation triangle. This
project is co funded by European Social Fund
through The Sectoral Operational Program
for Human Resources Development 2007-
2013, coordinated by The Bucharest Acade-
my of Economic Studies.
References
[1] S. Luján-Mora and J . Trujillo, “Extend-
ing UML for Multidimensional Model-
ling,” Proceedings of the 5th Interna-
tional Conference on the Unified Model-
ling Language, Germany, 2002.
[2] I. Lungu and A. Bara, Executive Infor-
mation Systems, ASE Printing House,
Bucharest, 2007.
[3] L. Moss and S. Atre, Business Intelli-
gence Roadmap – The complete project
lifecycle for decision-support applica-
tions, Addison-Wesley, 2004.
[4] A. R. Bologa, R. Bologa and A. Bara,
“Technology vs Business Needs in Busi-
ness Intelligence Projects,” Proceedings
of the International Conference on e-
Business (ICE-B 2008), 26-29 J uly, 2008,
Porto, Portugal.
[5] V. Diaconita, I. Botha, A. Bâra, I. Lungu
and M. Velicanu, “Two Integration Fla-
vors in Public Institutions,” WSEAS
Transactions on Information Science and
Applications, May 2008.
[6] L. Y. Liang and R. Miranda, “Dash-
boards and Scorecards: Executive Infor-
mation Systems for the Public Sector,”
Government Finance Review, December
2001.
[7] M. Z. Elbashir, P. A. Collier and M. J .
Davern, “Measuring the effects of busi-
ness intelligence systems - The relation-
ship between business process and orga-
nizational performance,” International
Journal of Accounting Information Sys-
tems, pp. 135–153, 2008.
[8] J . Hesford and K. Antia, A process-
oriented view of competitive intelligence
and its impact on organizational perfor-
mance, Working Paper. Richard Ivey
School of Business, The University of
Western Ontario, 2006.
[9] M. Edwards, “Best practices and leader-
ship in data warehousing award winners
2002,” Bus Intell J, Vol. 7, No. 4, 2002.
[10] A. Lönnqvist and V. Pirttimäki, “The
measurement of business intelligence,”
Information Systems Management, Vol.
23, No. 1, pp. 32–40, 2006.
[11] J . Marin and A. Poulter, “Dissemina-
tion of competitive intelligence,” Journal
of Information Science, Vol. 30, No. 2,
pp. 165–80, 2004.
[12] L. Davison, “Measuring competitive in-
telligence effectiveness: Insights from
the advertising industry,” Competitive
Intelligence Review, Vol. 12, No. 4, 2001,
pp. 25–38.
[13] Y. C. Lu and H. Cheng, Automated op-
timal equity portfolios discovery in a fi-
nancial knowledge management system,
computational intelligence in economics
and finance, Springer-Verlag, 2003, pp.
Informatica Economic? vol. 13, no. 4/2009 107
387–402.
[14] Z. Wang, Business intelligence, DrMas-
ter Culture Limited Company, 2005.
[15] T. Davenport, “Competing on Analyt-
ics,” Harvard Business Review, 2005.
[16] I. Lungu, M. Velicanu, A. Bara, V. Di-
aconita and I. Botha, “Portal based sys-
tem integration – foundation for decision
support,” Economic Computation and
Economic Cybernetics Studies and Re-
search Journal, Vol. 43, No. 1, 2009, pp.
123-135.
Adela BÂRA is a Lecturer at the Economic Informatics Department at the
Faculty of Cybernetics, Statistics and Economic Informatics from the Acad-
emy of Economic Studies of Bucharest. She has graduated the Faculty of
Economic Cybernetics in 2002, holds a PhD diploma in Economics from
2007. She is the author of 7 books in the domain of economic informatics,
over 40 published scientific papers and articles (among which over 20 articles
are indexed in international databases, ISI proceedings, SCOPUS and 2 of
them are ISI indexed). She participated as team member in 3 research projects that have been
financed from national research programs. She is a member of INFOREC professional associ-
ation. From May 2009, she is the director of the Oracle Excellence Centre in the university,
responsible for the implementation of the Oracle Academy Initiative program. Domains of
competence: Database systems, Data warehouses, OLAP and Business Intelligence, Executive
Information Systems, Decision Support Systems, Data Mining.
Iuliana BOTHA is an Assistant Lecturer at the Economic Informatics De-
partment at the Faculty of Cybernetics, Statistics and Economic Informatics
from the Academy of Economic Studies of Bucharest. She has graduated the
Faculty of Cybernetics, Statistics and Economic Informatics in 2006 and the
Databases for Business Support master program organized by the Academy
of Economic Studies of Bucharest in 2008. Currently, she is a PhD student in
the field of Economic Informatics at the Academy of Economic Studies. She
is co-author of two books, 13 published scientific papers and articles (among which two pa-
pers are ISI indexed and another 5 are included in international databases). She participated as
team member in 3 research projects that have been financed from national research programs.
From 2007, she is the scientific secretary of the master program Databases for Business Sup-
port and she is also a member of INFOREC professional association. Her scientific fields of
interest include: Database Systems, Design of Economic Information Systems, Grid Compu-
ting, e-Learning Technologies.
Vlad DIACONI?A is an Assistant Lecturer at the Economic Informatics
Department at the Faculty of Cybernetics, Statistics and Economic Informat-
ics from the Academy of Economic Studies of Bucharest. He has graduated
the faculty at which he is now teaching in 2005 and studies for his PhD in the
field of Cybernetics and Statistics. He is the co-author of 2 books in the do-
main of economic informatics, 2 articles in ISI journals, 4 articles in Scopus
journals, 4 articles in ISI proceedings, 4 papers in B+journals and 6 papers in
the proceedings of international conferences. He participated as team member in 3 research
projects that have been financed from national research programs. He is a member of INFO-
REC professional association. Domains of competence: Database systems, Data warehouses,
OLAP and Business Intelligence, Integrated Systems, SOA.
108 Informatica Economic? vol. 13, no. 4/2009
Ion LUNGU is a Professor at the Economic Informatics Department at the
Faculty of Cybernetics, Statistics and Economic Informatics from the Acad-
emy of Economic Studies of Bucharest. He has graduated the Faculty of Eco-
nomic Cybernetics in 1974, holds a PhD diploma in Economics from 1983
and, starting with 1999 is a PhD coordinator in the field of Economic Infor-
matics. He is the author of 22 books in the domain of economic informatics,
57 published articles (among which 2 articles ISI indexed) and 39 scientific
papers published in conferences proceedings (among which 5 papers ISI indexed and 15 in-
cluded in international databases). He participated (as director or as team member) in more
than 20 research projects that have been financed from national research programs. He is a
CNCSIS expert evaluator and member of the scientific board for the ISI indexed journal Eco-
nomic Computation and Economic Cybernetics Studies and Research. He is also a member of
INFOREC professional association and honorific member of Economic Independence aca-
demic association. In 2005 he founded the master program Databases for Business Support
(classic and online), who’s manager he is. His fields of interest include: Databases, Design of
Economic Information Systems, Database Management Systems, Decision Support Systems,
Executive Information Systems.
Anda VELICANU has graduated the Faculty of Economic Cybernetics, Sta-
tistics and Informatics of the Bucharest Academy of Economic Studies, in
2008. She is a PhD student in the field of Economic Informatics at the Acad-
emy of Economic Studies and since J anuary 2009, she is a Pre-Assistant
Lecturer. She teaches Database, Database Management Systems and Eco-
nomic Informatics seminars at the following faculties: Economic Cybernetics,
Statistics and Informatics, Commerce, Marketing and International Business
and Economics. Her research activity can be observed in the following achievements: 5 dip-
lomas, 2 scientific awards, 3 proceedings, 2 articles published in scientific reviews, 1 research
contract, 1 book and 1 research grant. She is a member of INFOREC professional association.
Her scientific fields of interest include: Databases, Database Management Systems, Pro-
gramming, Information Systems.
Manole VELICANU is a Professor at the Economic Informatics Department
at the Faculty of Cybernetics, Statistics and Economic Informatics from the
Academy of Economic Studies of Bucharest. He has graduated the Faculty of
Economic Cybernetics in 1976, holds a PhD diploma in Economics from
1994 and starting with 2002 he is a PhD coordinator in the field of Economic
Informatics. He is the author of 18 books in the domain of economic infor-
matics, 64 published articles (among which 2 articles ISI indexed), 55 scien-
tific papers published in conferences proceedings (among which 5 papers ISI indexed and 7
included in international databases) and 36 scientific papers presented at conferences, but un-
published. He participated (as director or as team member) in more than 40 research projects
that have been financed from national research programs. He is a member of INFOREC pro-
fessional association, a CNCSIS expert evaluator and a MCT expert evaluator for the program
Cercetare de Excelenta - CEEX (from 2006). From 2005 he is co-manager of the master pro-
gram Databases for Business Support. His fields of interest include: Databases, Design of
Economic Information Systems, Database Management Systems, Artificial Intelligence, Pro-
gramming languages.
doc_385719315.pdf