On the Generation of Dynamic Business Indicators
Fábio Alexandre Pereira dos Santos, Rui César das Neves and Joaquim Belo Filipe
Instituto Politécnico de Setúbal, Escola Superior de Tecnologia de Setúbal, Setúbal, Portugal
Keywords: Dynamic Generation of Data, Business Intelligence, OLAP, Data Representation, Infographics, Entity
Framework.
Abstract: While information is rapidly gaining relevance to the organizations, systems that help companies analyse
that information need to improve their effectiveness in several layers. We address, in our on-going research
work, mainly the presentation layer and business layer of software systems development. The aim of this
position paper is to discuss how to develop a system that is as flexible and configurable as possible, which
allows multiple methods of analysing and visualizing the relevant data to an organization, allowing them to
define certain views on business data with appropriate graphics. One of the values from this system is the
domain of technology in a controlled environment, because it helps solving a specific problem, since it is
both a generic tool, autonomously and with a high degree of adaptability. Flexibility will go hand-in-hand
with ensuring the consistence of the data model, i.e. the data being analyzed by an user must respect
syntactic and semantic constraints when relations to each other created in the organization’s logic. This
feature prevents the user to attempt to connect data that aren’t related. Based on the existing data types
described on metadata in the data model, the system provides the users with a list of possible graphical
representations for the selected information type. This list will be filtered in order to allow the user to select
only graphical representation types that are appropriate to the selected data types. This is an innovative
feature, in the sense that the system constrains the selection of the visualization elements thus avoiding
potential conceptual errors.
1 INTRODUCTION
In recent times an increasing attention has been
given to take advantage of information and
knowledge in organizations in order to gain
competitive advantage.
Thus, the information about organizations is now
an essential component of support for their
operations, since it allows creating internal and
external favourable conditions to reduce costs and to
provide innovative services. It also facilitates the
construction of knowledge that is required to plan
and implement solutions for problems and
challenges that arise everyday on an organization.
In organizations, are agents there are responsible
for decision making and in this process they require
rigorous tools to analyse the data and further
improve the performance of the organization. Those
agents need to access and process different types of
relevant data.
Unfortunately the development of information
systems is, usually, more concern in the
development of the information input areas, creating
system that collects information and guarantees the
information coherence and leaving the outputs areas
with limited capabilities.
The main goal is the development of a tool that
could be connected, in a standard way,
independently from the application domain, to an
already developed system which gives information
extraction capabilities to the end user. These
capabilities will allow the drill down in all
information available, maintaining the coherence
and relations defined in the input area.
With this in mind, the system presented in this
position paper is a tool to help companies mining
their own data and obtain new relevant information,
which is not accessible from the traditional
information systems.
Actually there are a lot of systems doing similar
things, but most of them are not as specific and
configurable as desired. The systems that are
configurable often have features that most
organizations do not need.
390
Alexandre Pereira dos Santos F., César das Neves R. and Belo Filipe J..
On the Generation of Dynamic Business Indicators.
DOI: 10.5220/0004199603900394
In Proceedings of the International Conference on Knowledge Discovery and Information Retrieval (KDIR-2012), pages 390-394
ISBN: 978-989-8565-29-7
Copyright
c
2012 SCITEPRESS (Science and Technology Publications, Lda.)
2 IMPORTANT TERMS
AND SYSTEM MODEL
In this section, it is important to clarify a few critical
terms that are needed to set the stage for everything
else that follows.
The first term is entity, which is defined as a
concept in the domain of application from which a
data type is defined, where data type is an element
that represents the structure of a top-level concept
and is a template for instances of entity types in a
system.
Another important term is indicator, which is
defined as a set of data belonging to one or more
entities which are related to each other in the data
model. It’s not only necessary to be related, but also
that they have the same meaning in the
organization’s business logic.
An important aspect of the system that we will
implement based on these ideas is its generic and
configurable structure, independent of the
application domain and requiring only a set of
rigorously defined terms in order to allow navigation
through all information.
Thus, the aim is to develop a web-based system
that allows structured access to information,
according to the organization’s business logic. This
gives to the end user the opportunity to do queries
using a configurable and appropriate interface. The
integrity of these queries will be guaranteed through
the modeling process required by ORM (Object/
Relational Mapping), because this, ORM, will
guarantee that the end user has restrictions on data
access, avoiding mistakes, unlike conventional
systems in which the ORM is applied on developer
side.
Based on this, the system is built around the
approach of the Entity Framework (EF) from
Microsoft, ADO.NET. There are other ORM
technologies available in the market, but we choose
this one because it comes from a major software
supplier, is the ORM embedded in.NET Framework,
and is the approach that we are studying.
It is also important to mention that changes in the
data model shouldn’t force the rewrite of the whole
system and should allow the reuse of all available
information.
The idea is to provide to end users a list of
possible graphic representation forms for an
indicator, because the aim is to constrain possible
actions by the user, thereby reducing the probability
of mistakes on selecting a graphical representation.
We select that kind of representation, although
there are several ways to display and process these
types of data, but the most simple is through
graphical representation, being this conclusion given
by (Tufte, 2001) “(…) graphics are instruments for
reasoning about quantitative information. Often the
most effective way to describe, explore, and
summarize a set of numbers – even a very large set –
is to look at pictures of those numbers. Furthermore,
of all methods for analyzing and communicating
statistical information, well-designed data graphics
are usually the simplest and at the same time the
most powerful”.
These systems, which help decision-making,
must have the simplest forms of data representation,
i.e., it should be able to easily illustrate the graphical
representation of sums, dates and tables.
It is also expected to illustrate facts that have
occurred in time, i.e., present time series, for a given
data, because with only one dimension this kind of
representation allows to display time series, with an
appropriate scale that can’t be achieved with another
type of graphical representation.
Another important feature is the presentation of
indicators regarding space, because organizations
typically have data that is related to locations /
regions and are of great importance to the
organization’s business model.
The graphical presentation is one of the key
points in this system, because according to (Tufte,
2001) “(…) No doubt some graphics do distort the
underlying data, making it hard for the viewer to
learn the truth”. Thus, it is important that the system
can filter out these types of representations,
according to the types of data that compose the
indicator; this feature will be described in section 3
of this paper.
The system that we are implementing is clearly
placed in the scope of the so-called OLAP, i.e On-
Line Analytical Processing.
OLAP provides the ability to analyse different
aspects of information in a fast and dynamic way,
where large volume of data are contained within a
data warehouse according to (Thomsen, 2002) “(…)
OLAP is meant to contrast with OLTP (On-Line
Transaction Processing). The key aspects are that
OLAP is analysis-based and decision-oriented”.
Until recently these systems were known as DSS
(Decision Support Systems), but now it is common
to refer to them as Business Intelligence (BI)
systems, which according to (Rud, 2009), “(…) BI is
defined as the ability for an organization to take all
its capabilities and convert them into knowledge
(…)”.
The concept of BI is relatively recent and BI
systems are usually composed of a set of tools that
enable report generation and allow users to extract
OntheGenerationofDynamicBusinessIndicators
391
useful information from the stored data.
BI systems have also a set of tasks associated,
which according to (Cardoso, 2011) “(…) can be set
into four groups:
Make predictions based on historical data,
based on past performance and current;
Creation of scenarios that demonstrate the
impact of changes to existing variables;
Allow ad-hoc access to data;
Analyse in detail the organization, thereby
ensuring a deeper knowledge about the
same.”
With the sets of tasks and features shown above,
the traditional systems could work connected to the
BI systems yet independently, allowing to apply
specific techniques, which according to (Santos and
Ramos, 2009) “(…) BI systems have implemented
the functionality, scalability and security of existing
systems which manage the database for building
data warehouses that are analysed with techniques
for OLAP, Data Mining and Query Report”.
The system being developed cannot be
considered a BI tool as it doesn’t cover the first two
groups identified by the (Cardoso, 2011), just
covering the third and fourth points - allow ad-hoc
access data and in detail analysis of the organization.
At the data source level, the system needs a
model that contains information about the present
data and the relationships between objects that
compose this data.
The next section will describe forms of
presentation for indicators.
3 DATA AND DATA TYPE
REPRESENTATION
With the information available from the EF, the
system must advise users what types of common
graphical representation can be used to see the data.
After analyzing the major database vendors, namely
Oracle, SQL Server and MySQL, it can be
concluded that all kinds of data may be grouped in 4
groups, which we call Numeric, Date and Time,
Strings and Other Types of data that we don’t want
to represent, like “XML”, “Cursor”, “BLOB”,
“BFILE” types, etc..
The most common forms of representation of
these groups of information are “Column”, “Lines”,
“Pie”, “Bar” and “Scatter”.
The scatter plots (X, Y) usually shows values in
both axes X and Y, while the other graphs show
usually in the Y axis, using “categories” in the X
axis.
This information is very important because it
allows the system to tell the user which data
representation can be selected. Better yet, according
to the data types of the selected information, we can
tell the user what kind of graphical representation
shouldn’t be selected, avoiding the user to select
forms of graphic representation which may not be
suitable to the selected data.
Initially we want to use simple graphics, which
can be represented by 2D representation, because we
need to find an efficient way to help users avoiding
errors on select data representation.
Those simpler graphic representation forms can
be divided into two groups, depending on the data
types that we are working on. Those groups are
Categories vs. Values and Values vs. Values, as we
can see on
Table 1, where categories are like objects
and values are like the object values. The second one
is a traditional graph, where values for both axes are
available.
Table 1: Forms Representation.
The idea of this mapping, between the existing
data types and forms of graphical representation, is
to show that there is dependence between the
existing data types and the ways to represent them.
According to the table above, there are types of data
that cannot represent certain forms of information
and that when displayed on screen are shown in the
wrong way and therefore lose their meaning.
4 DEVELOPMENT
In the following sections, the main focus is to
describe how we want to solve the problem,
explaining the specifications and highlighting some
implementation details.
KDIR2012-InternationalConferenceonKnowledgeDiscoveryandInformationRetrieval
392
4.1 Requirements Analysis
The requirements for the project are divided into two
different groups: technical and user interface.
4.1.1 Technical Requirements
Regarding technical requirements, it is expected that
the system is capable of accessing information types
at runtime and validate new data relations that the
user may wish to create.
In the EF, the conceptual model, storage model
and the mapping between the two is defined on an
.edmx file. This file is updated when either the
database or the model changes. According
(Corporation, Modeling and Mapping, 2011), “the
EDM Generator, which is included with .Net
Framework, generates the .csdl, .ssdl and .msl files
from an existing data source”.
These files, XML-based, describe the conceptual
model, storage model, and mapping, are known as
metadata.
This means accessing the metadata which was
represented in .csdl and .ssdl files and was loaded
into instances of the
“System.Data.Metadata.Edm.EdmItemCollection”
and
“System.Data.Metadata.Edm.StoreItemCollection”
classes, which are accessible by using methods in
the
“System.Data.Metadata.Edm.MetadataWorkspace”
class.
These instances allow the system to understand
the structure of data, thus, allowing navigation
through all information to create query’s
dynamically as well as an appropriate representation
for the user.
4.2 User Interface
The second group of requirements defines the user
interface: users will use the application to analyse
data which originated from the data source.
The system should enable its users to create and
display dynamically a set of indicators through
graphical representation, which can be customized
for each user.
The user should be able to select or even create
data visualization models at runtime, in a highly
configurable way: the user could add a new graphic
representation for an indicator, he/she could also
remove certain graphical representation forms or
modify parameters of others.
If the user wants to create a new graphical form
to represent the indicator, user should select which
properties from entities wants to relate and after that
the system will process the selected data. After that,
the system will show to user the output. If the user
wishes to add more constraints to this process, may
at any time add it.
This process, the selection of properties from
entities, should operate in a cycle, and this cycle is
summarized in Figure 1, allowing refining the
output, which will be shown on the user interface.
Figure 1: Operating Workflow.
As the user can have a large amount of graphics
and indicators representation, the idea is to give the
ability for to personalize her/his interface, deciding
the order in which the content should appear on the
page, for example. The user should be able to
minimize, move items and presentation forms
around the page as if it were a design surface and to
remove items as needed.
The system will be required to maintain the user
page customization, in order to avoid repeating the
visualization settings each time the user visits the
system. This can be done using Web Parts, a
technology that provides an appropriate way to build
a modular Web Site that can be customized, with
dynamic settings, on a per-user basis. This
customization is provided by one provider model –
Personalization Provider, and according to some
authors (Evjen et al., 2010) “(…) this provider
makes associations between the end user viewing the
application and any data points stored centrally that
are specific to that user”, giving exactly what we
need.
The approach used in the proposed
implementation follows a similar approach to the
traditional providers, according to the same source
(Evjen et al., 2010) ”a provider is an object that
allows for programmatic access to data stores,
processes and more”, and in this systems’ context
means the independence of a data model.
OntheGenerationofDynamicBusinessIndicators
393
5 CONCLUSIONS AND FUTURE
WORK
In this position paper, we presented a comprehensive
set of criteria for the development of a data
visualization system. This system is based on
Microsoft EF and it has, explicitly or implicitly, a
constrained conceptual data model that describes the
various elements of the problem domain.
The data model represents concepts and also the
relationships between concepts, constraints, and so
on.
Since currently most applications are written on
top of relational databases, they will have to deal
with data represented in a relational form. The
programming paradigm is typically some form of
Object-Oriented Programming (OOP) including
features such as data abstraction, encapsulation and
inheritance, and these features are fundamental to
the flexibility that is intended to characterize this
system.
Typically a higher-level conceptual model is
used during the design phase, and that model is not
directly executable, so its need to be translated into a
relational form and applied to a logical database
schema and to the application code.
With this flexibility guaranteed by EF, our
system differs from many other BI systems on this
point as well, because the great majority run on top
of traditional databases, i.e. other systems are based
on multiple data sources but ours just needs to
connect to a schema from EF.
Another feature of our system is the ability to
help users to visualize the information that was
identified as relevant business indicators, based on
the selection and combination of data types or
dynamic relations and constraining the choice of
possible graphical representations.
This feature is also important because it avoids
user errors when selecting the graphical
representation forms for the selected information.
The system that we are developing will use an
approach like the providers approach because, as we
mentioned above, this feature allows the user to
access data via a web browser, in readable form,
without the need to use a complex system.
It is intended that the system will evolve to an
interface that allows the user to create a
configuration of graphical representation forms, as
well as the possibility of adding new types of
graphical representation and mapping them to
existing data types.
ACKNOWLEDGEMENTS
We would like to thank the Polytechnic Institute of
Setúbal, School of Technology of Setúbal, for
supporting the research work reflected in this paper,
presented at KDIR 2012 in the scope of the RETE
project.
REFERENCES
Cardoso, E. (2011, Setembro). Introduction to Business
Intelligence. Lisboa, Lisboa, Portugal.
Corporation, M. (2006, June). The ADO.NET Entity
Framework Overview. Retrieved 04 2012, 04, from
MSDN: http://msdn.microsoft.com/en-
us/library/aa697427(v=vs.80).aspx
Corporation, M. (2011, 8 29). Modeling and Mapping.
Retrieved 5 5, 2012, from MSDN:
http://msdn.microsoft.com/en-us/library/bb896343
Evjen, B., Hanselman, S., & Rader, D. (2010). ASP.NET 4
in C# and VB. Canada: Wrox.
Inmon, W. (2002). Building the Data Warehouse. Willey.
Rud, O. P. (2009). Business Intelligence Success Factors:
Tools for Aligning Your Business in the Global
Economy. Hoboken, New Jersey: Wiley & Sons.
Santos, M. Y., & Ramos, I. (2009). Business Intelligence.
FCA.
Thomsen, E. (2002). OLAP Solutions - Second Edition.
Willey.
Tufte, E. R. (2001). The Visual Display of Quantitative
Information 2ª Edition. Graphics Press LLC.
KDIR2012-InternationalConferenceonKnowledgeDiscoveryandInformationRetrieval
394
