17th International Conference on
Database and Expert Systems Applications
September 4-8, 2006
A Spatial Dashboard: analyzing business performance measurements
using the spatio-temporal context
Gabriel Pestana1, Miguel Mira da Silva2, Mário Gomes2
INESC-ID/IST, Rua Alves Redol, 9,
1000-029 Lisboa, Portugal
IST, Rua Alves Redol, 9,
1000-029 Lisboa, Portugal
The advent of mobile computing and location-based services has changed the way decision-
makers analyze business. This growth in spatial data collection has shift decision-makers
attention to the benefits of considering the spatial dimension as an additional parameter in
business performance management. In this paper we present an innovative approach, named
Spatial Dashboard, which describes decision-makers information requirements to analyze
business performance using the spatio-temporal context. The Spatial Dashboard makes use of
Treemaps, a visualization technique for presenting Key Performance Indicators values on two
dimensional displays. Historical data analysis are accomplished using Timeboxes, a rectan-
gular graphical user interfaces to specify query constraints on time series data sets. Using a
practical example in the field of airport traffic control, we evaluate the implementation of the
With the advent of mobile computing and location-based processes, the spatial and tempo-
ral dimensions are becoming a corner stone for decision-makers to analyze business data .
This growth in spatial data collection enabled the decision process to take a deeper focus on
the representation and propagation of continuous changes at different levels of abstraction
However, during the analysis process, business data are frequently striped away from the
context of the events captured, including time, location, and other contextual elements. Con-
sequently, the decision process does not take into account context-awareness parameters such
as ongoing changes to spatial data and business rules.
Therefore, we propose to measure business performance from a spatial and temporal point
of view by extending the traditional Key Performance Indicators (KPI) framework  to the
spatial domain. These KPI can be a physical measure such as Sales, a calculated measure such
as Profit, or a calculation that is defined within the KPI.
Using the Balanced Scorecard (BSC) methodology , KPI are grouped into perspectives
(e.g., Customer, Processes, Innovation, Finance). The contribution of each perspective is fur-
ther weighted to determine the corporate performance.
Decision-makers frequently rely on the Data Warehouse multidimensional data structure as
an important platform for data analysis . The multidimensional model is usually repre-
sented as a star/snowflake schema , made up of a set of (fully denormalized) dimension
tables, one for each dimension of analysis, and a fact table that stores the set of measures and
whose primary key is obtained by composing the foreign keys referencing the dimension
tables. A dimension is usually organized in hierarchies supporting different levels of data
aggregation as well as multiple inheritances.
The proposed model, called Spatial Dashboard, aids decision-makers to explore the hierar-
chical nature of multidimensional data structures. It is oriented to the modeling of gradual
changes in the spatial and temporal domain. The approach is based on an activity and process-
based model that allows for the representation and propagation of continuous changes at dif-
ferent levels of abstraction. This observation of changes is adequate for coping with business
volatility and, hence, dynamic-evolution at the business domain level .
The Spatial Dashboard provides two types of indicators: basic indicators (meaning that the
value of the indicator is determined from measures stored in the fact table), and derived indi-
cators (meaning that the value of the indicator is determined based on an expression over
other basic or derived indicators). A derived indicator is usually associated to a KPI and basic
indicators to business performance indicators (PI). A KPI can also be defined by a basic indi-
cator if it is consider as a critical measure for business performance measure. We will refer
both concepts (KPI and PI) just as PI and differentiate them whenever required.
The Spatial Dashboard provides an abstract visual representation of PI in a single screen
display that can be grouped and arranged by perspectives or drilled-down to line-of-business.
As in , each indicator is represented by a rectangle with the size proportional to its “signifi-
cant impact” in corporate performance. The colour is propositional to the cost variance. Busi-
ness performances below target values are represented in red, in yellow all the deviations
within a default tolerance and green when business performances are on or above target val-
The remaining of the paper is structured as follows. Section 2 presents the modeling back-
ground and related work. Section 3 introduces the conceptual model approach. Section 4 de-
scribes the architecture of the model and applies those concepts. Section 5 describes in detail
the Semantic Manager component. Finally Section 6 draws some conclusions.
2. Related Work
The importance of understanding context-aware changes, as well as, the effects and causes
in spatial planning is a well discussed subject within the research community [2, 10, 17, 19].
Change and related constructs such as events and processes are key concepts to establish ap-
propriate representation for change based on ongoing events and in particular to conceptual
modeling of dynamic systems.
Several conceptual and logical models have been proposed [6, 10, 20] to represent changes,
processes and events in Geographic Information Systems. However, the proposed models
fetch to tackle high-order tasks such as finding the rules of change according to different ab-
straction levels and enabling multiple representations of the facts according to distinct per-
Our model incorporates the combined effect of roles and hierarchies to enable information
integration across different abstraction levels. The model extends the concept of multi-
representation of geographic data, namely the SOLAP and VUEL concepts , to the busi-
ness performance management domain for multiple points of views and abstraction levels.
Pag. 2 - 12
In  the benefits of a geographic visualization are also discussed. The proposed method
is a map-centered implementation of the analytic hierarchy process (AHP), however with a
strong operational oriented approach. The work in  presents some common aspects but
they do not have an architectural approach to handle spatio-temporal evolutions of the ob-
In the literature the ability for decision-makers to monitor business performance on an on-
going basis , enabling them to make conditional predictions based on alternative hypo-
thetical scenario is not a well explored subject . In our opinion a good approach to this
subject is to consider space and time as conceptually equivalent and treated as such form the
modeling point of view. Some authors  even say that Time should be considered as a
framework in which change is possible.
When dealing with conventional planning, the BSC appears to be a well known working
methodology  which can be extended to support the spatio-temporal modeling process for
quantifying performance metrics. Nevertheless, while providing a vast array of individual
quantitative indicators, the BSC framework does not provide consolidated performance val-
ues, either for the individual perspectives or for their consolidation .
Many BSC systems provide companies with the capabilities to graphically display focused
groups of metrics concerning the process. These displays are called, variously, portals,
dashboards, cockpits or scorecards .
“Rows and columns” are necessary, but when decision-makers need to know if there is a
problem at a glance, different visual cues should be used (e.g., speedometers, thermometers,
stoplights, etc.) in describing whether business performance status is better or worse. The use
of gaugers has some drawback because they overcrowd the screen display when decision-
makers need to monitor the business status at a fine level of granularity. A gauger also does
not provide any visual clue about the hierarchy of indicators ordered by their contribu-
tion/impact to each perspective, abstraction level or corporate performance. The proposed
model uses TreeMap  functionalities to represent business performance status.
To the best of our knowledge there is no conceptual modeling approach that addresses lo-
cation-awareness in business processes in the sense that they are suitable for multi-
representations of PI for different perspectives and abstraction levels. Existing methods fail to
address requirements for multidimensional spatial problem solving, as they are inflexible and
often domain-specific with no support to how business indicators relate and evolve using the
3. The Spatial Dashboard
The Spatial Dashboard model has emerged from the BSC framework . It corresponds to
a spatially enabled multi-perspective approach within a single consolidated view that inte-
grates functionalities traditionally split between diverse technological solutions, namely,
Geographic Information Systems, Data Warehousing, On Line Analytical Processing (OLAP)
analysis, performance measuring and modeling tools.
Our approach for analyzing business performances is though the geospatial correlation of
PI with the corresponding business processes. This is an activity oriented approach, i.e., for
each activity within a business process, we identify the spatio-temporal primitives that apply.
The Spatial Dashboard lies at the intersection of decision support and “right time”, multi-
dimensional business data analysis. The term “right time” means that the information is avail-
able when it is required. With right-time visibility and control of IT operations, events can be
correlated to KPI (time-critical business parameters) for meaningful and actionable insight for
strategic, tactical and operational decision-makers to readily intervene whenever required.
Pag. 3 - 12
Besides of monitoring the progress of activities, decision-makers also have to be timely noti-
fied of any conflict/infringement.
The Spatial Dashboard requires indicators to be constantly fed so that they may be avail-
able at the right-time, and at the proper decision level. Hence, we define the data latency as
the interval between the time an event occurs and the time it is perceived by the user.
We use dataflow stream features, namely the Proactive Cache provided by the Analyses
Services of Microsoft SQL Server 2005 , to accomplish for shorter information latency.
The Analyses Services capability enables decision-makers to analyze data across an array of
operational systems without persisting the data. The data can flow directly into an Analysis
Services partition from the Data Transformation Services (DTS) package pipeline, without
intermediate storage. This scenario reduces the latency (and storage cost) of analytical data.
In addition to right-time information, decision-makers are also interested in trends .
They must know whether the organization’s performance is improving or worsening, when-
where-why this is happening, and which PI needs immediate attention.
The Spatial Dashboard viewer provides a set of graphical functionalities to simplify the
data analysis process using the spatio-temporal context. Figure 1 shows some of Spatial
Dashboard main components, namely: Trend Icon, TreeMap rectangles, sliders, TimeBox,
MapViewer and Semantic Manager.
Fig. 1 – The Sketch of the Spatial Dashboard second level.
Trend Icon: it represents the Corporate Trend (see Figure 1, icon at the top left corner).
Trends are based on scores and are determined by comparing actual values to target values
over time. Trends are usually represented as an up arrow ( ) for performance improvement, a
down arrow ( ) for performance worsening, and a horizontal bar (=) to indicate either no
change or a change that is within the trend tolerance percentage.
TreeMap rectangles: at the Spatial Dashboard viewer, scores are represented graphically
as rectangles using the TreeMap  functionalities. The size of a rectangle is proportional to
Pag. 4 - 12
the contribution/impact of each metric. By clicking on a metric, the corresponding descriptive
attributes are presented at the “Attributes of the Indicator” viewer.
Sliders: decision-makers may navigate through the metrics using depth sliders to control
the level of detail, i.e., to drill-down into more granular data structures. A time slider is also
provided to enable decision-makers to visualize (rollback and forward) the sequence of events
within the exact spatial and temporal context they occurred.
TimeBox: used to pose queries over a set of entities with one or more time-varying attrib-
utes. Timeboxes are rectangular query regions drawn directly on a two-dimensional display.
The extent of the timebox on the time (x) axis specifies the time period of interest, while the
extent on the value (y) axis specifies a constraint on the range of values of interest in the
given time period. In Figure 1, the extent of the blue rectangle – TimeBox, defines a Semester
MapViewer: this component enables decision-makers to analyze the PI using a map-based
form. Through a map-based layout, decision-makers may benefit from spatio-temporal func-
tionalities to better exploit and analyze the data. We use stereotypes called pictograms  to
signal spatially enabled PI. In Section 5 we refer how to associate spatiality to a PI so that it
may be analyzed in a map-base form.
Semantic Manager: this component enables decision-makers to specify the business indi-
cators within a specific context and having rules or expressions to evaluate the current
status/trend for the measure.
The Spatial Dashboard viewer provides two main screen displays. The fist screen presents
a consolidated view of each perspective with the size of the rectangles proportional to the
contribution of each perspective to the corporate performance score. The second screen level
presents the overall hierarchy structure of the performance Indicators. All the rectangles be-
low the chosen hierarchy level are aggregated and represented by a single rectangle. Accord-
ing to  the size of the aggregated rectangle is the sum of the size of the items below in the
chosen hierarchy level.
Each rectangle also provides information about the metric scores. Scores represent per-
formance numerically, i.e., whether the organization is on or off target and how far on or off
target. Each score is calculated from the actual, target, and tolerance values of the metric, plus
the performance pattern (weight) for the metric type. The score parameters are presented in
equation (1) were the weight of each performance indicator is expressed by WPI. The consoli-
dated view of all PI for each perspective is calculated using expression (2) and the Corporate
Performance score is determined as in (3).
C = ∑ PI i ∗ W PI i where C refers to the perspective score (2)
CP = (C * WC + IP * WIP + I * WI + F * WF) / Total Weight
Pag. 5 - 12
Because the score is calculated from all three values, it can go down even when the actual
value goes up. Like Trends, status indicators are also based on scores and represent perform-
ance graphically. Figure 2 present the three minimum states.
Fig. 2 – The three status states. Green - on or better than target, Yellow - off target
but within tolerance, or Red - off target and not within tolerance, and Minimum - the
lowest score to assign.
In addition to the functional components presented before, a set of alert messages (with dif-
ferent severity levels) notify decision-makers whenever a metric value approaches or under-
exceeds a default threshold. Alert messages are activated based on the PI status and business
rules which are structured hierarchically.
Figure 3 shows how the Spatial Dashboard hierarchical structure (left) is being imple-
mented using the analytical hierarchy process (AHP) method  (right). The AHP method
decomposes the decision problem in an attribute hierarchy, rooted at a single attribute repre-
senting the overall (corporate) score.
Instead of comparing all possible attributes at once, the establishment of such a hierarchy
reduces the cognitive load of importance weighting to only those criteria that fall under the
same group attribute.
Fig. 3 – The abstraction levels of the Spatial Dashboard model.
We started the implementation with the AHP method due to its simplicity and efficiency to
aggregate data for all levels of the hierarchy structure. For small problems AHP method is
often the starting point algorithm for data exploration and predictive modeling of both dis-
crete and continuous attributes.
Other, more complex methods, such as fuzzy logic, shall be implemented only after the
model has been fully tested using the AHP method. The fuzzy logic method is more complex
because it deals with impression and ambiguity in the input data (e.g., attributes values and
decision maker’s preferences).
In most situations the hierarchy structure and metric weights are defined by business ex-
perts. Some rules have also to be carefully defined because they capture the logic of individ-
ual decision-making activities. These rules can take the form of a validation condition or a
more complex mathematical expression to provide consistency of behavior to support ongo-
Pag. 6 - 12
ing processes. We may conclude that business processes and rules are closely tied and in
many cases they require some level of human intervention .
On the other hand, business rules, metrics and the corresponding weights can change
quickly because of business dynamics. Therefore, it is important that the modeling process
properly maps the business processes and their relationships in order to find out the relevant
indicators and rules, and to determine where the data to compute them can be found.
4. The Architecture of the Model
In this Section we present an overview of the Spatial Dashboard architecture that is being
developed. It is a component-based development platform that brings together intrinsically
related parts of process-oriented information systems.
The focus of the proposed model is to monitor the activity flows, and location-aware in-
formation for constructing mutirepresentations of observed phenomena  (e.g., business
activities consisting of events that occur at a time and location) and to provide the appropriate
information representation for each level of the decision process (i.e., strategic, tactical and
operational). Figure 4 shows the building blocks of the Spatial Dashboard architecture to-
gether with the technology being used.
Fig. 4 - The Architecture of the Spatial Dashboard.
A short description of the main building blocks is presented next.
GUI: the Graphical User Interface is responsible for managing the interface between the
user and the rest of the components. The GIS kernel routines are responsible to handle desk-
top mapping functions.
Persistence Layer: this layer is responsible for storing the data, either spatial or non-
spatial. The multidimensional data structure of the database enables information standardiza-
tion . Data are stored using established database products such as SQL Server 2005.
BPM Engine: this component is responsible for modeling business processes using the
Pag. 7 - 12
SOLAP Engine: this component supports the exploration and analysis of spatial data. The
data exploration is achieved with operators such as drill-down, drill-up, drill-across and slice,
without requiring any technical expertise in database query language.
Scenario Manager: enables decision-makers to simulate and test what-if scenarios without
affecting the current values of the actual data. All the scenarios can be stored for future test-
ing. The level of understanding is tested against the ability to make conditional predictions
based on alternative hypothetical scenarios (e.g., answering to what-if questions).
Spatial Modeler: provides a graphical modeling interface for designing and implementing
geoprocessing models. The workflow diagram links together spatial operators and data, creat-
ing a visual representation of data and processes as on-screen objects or icons with connecting
lines that show model flow or order of execution.
Semantic Manager: makes the semantics visible. The ontology-driven concepts are repre-
sented as a hierarchical tree of the real world objects . This component also provides the
functionalities for decision-makers to create new PI.
Rule Engine: implements the decision-making logic. The Rule Engine is responsible to
automate business decisions, to structure loose informal business practices and policies, and
to apply sets of conditions that control specific business events.
The Spatial Dashboard model is centred on semantic modeling primitives that capture busi-
ness performance using the spatio-temporal context. Indeed, our purpose is to bring spatio-
temporal awareness to the business performance conceptual modeling level. The next section
will focus on the Semantic Manager component because it is where the context structure is
5. Semantic Manager
In this Section we describe how metric are created and how they are automaticaly assigned
to a semantic context.
As presented in Figure 5, to create new PI decision-makers have to specify a set of parame-
ters. They have to specify the level of abstraction (Strategic, Tactical or Operational), if the
indicator is a Benefit, Cost or On-target measure, if it is a basic or derived indicator and de-
fine the corresponding spatial parameters.
Pag. 8 - 12
Fig. 5 – The Semantic Manager Interface.
The representation of metrics in the form of a map is accomplished through a call to the
Spatial Modeler. This component enables decision-makers to create a spatial model for cop-
ing with business volatility and, hence, dynamic-evolution at the business domain level.
The spatial model correlates a set of spatio-temporal functions and converts input data into
an output map. The output map may have geographical notations such as Point, Line, and
Polygon features represented as pictograms such as , , or . Pictograms constitute a spa-
tially-aware language used to differentiate spatial form non-spatial indicators.
The spatiality of an indicator is automatically externalised to all the intervenient compo-
nents of the Spatial Dashboard architecture.
We use temporal pictograms  to model geometric evolutions of objects over time. Picto-
gram is used if the measure values or geometry are valid only for an instant and pictogram
if they are valid for longer periods. Selecting between or depends on the temporal
granularity defined for each measurement or descriptive geometric attribute.
Our geoprocessing workflow combines data with sets of logic (business rules) and spatio-
temporal manipulation functions to describe the observed phenomenon, such as airport traffic
flow for ground vehicles.
Considering the aeronautical sector as an example, a spatially enabled PI would be the av-
erage delay in assisting parking aircrafts or the number of restricted area incursions caused by
vehicles. When the observed events report an infringement or detach themselves from the
average values decision-makers may analyze the data using a map-based layout. For instance,
to backtrack the exact sequence in which activities were accomplished within the defined
The MapViewer Location-aware and time series data analysis functionalities can help deci-
sion-makers to identify some correlations on business performance measurements. Moreover,
since all business activities consist of events that occur at a time and location [9, 20], these
two contextual elements, time and location, are useful for organizing data about events.
External events can have a strong impact on business performance. Therefore, it is com-
mon for decision-makers to analyse the impact of external events in the organization needs,
namely in its internal business processes, activities and target values.
Pag. 9 - 12
Taking the airport safety field as an example, in 2004 Portugal held two international
sports events (the UEFA EURO2004 Football Champions League and UEFA Final Cup). As
a consequence the promotion of Portugal as a destination of interest was intensified, causing a
huge increase in the number of flights.
The increase on airport movements forced some changes to business processes and rules.
During the realization of the events a set of measurements registered abnormal scores,
namely, the number of Conflict/Infringements, Level of Safety and Security on the Movement
Area, Adequacy of the Service Quality, Competitive Cost Structure and Customer Satisfac-
The score variance on most of those metrics derived from changes to business processes
and rules. Some of them decreased the level of Safety and Security in the movement areas and
others increased airport operations efficiency and customer satisfaction (e.g., aircraft business
partners and companies).
After the sport events, a set of simulations were tested using the Spatial Dashboard proto-
type to enable decision-makers to test some scenarios. The aim was to analyze the traffic be-
havior for different alternative situations (e.g., business rules, increase of the Apron area,
changes to aircraft circulation paths and parking areas) and try to increase airport efficiency
and reduce the number of conflict/infringements.
The Spatial Dashboard prototype enabled decision-makers to do map-based data analyses
and to visualize the simulated movements (including the conflict/infringements alert mes-
sages) within the spatial context for each period. The map-based layout provided all the in-
formation needed for decision-makers to understand the scenarios with conflict/infringement
alert messages. It was also possible to analyze which correction action (what-if scenarios)
could increase corporate performance score. Some of the multi-performance outputs were
perceived by decision-makers just by looking to the MapViewer.
Safety issues were analyzed from different perspectives and abstraction levels (see Figure
6). At the left side, Figure 6 presents a high-level map-view (Strategic level) of the aggregated
Conflict/Infringement messages. At the right side, the map-view presents a drill-down view
into one of the critical areas (signaled with red circles on the left side). At this level of data
granularity it is possible to see the vehicles and aircrafts movements and the corresponding
protection areas on an ongoing base.
Fig. 6 – Monitoring conflict/infringements caused by vehicles and aircrafts move-
ments at different abstraction levels.
The tests had the unmatched capacity to provide answers to questions that are central to
safety, including the prevention of future incidents/accidents:
Pag. 10 - 12
• What is happening or when did it happen?
• Where are the areas that need the most attention? What do I need to do about it?
• Why is it happening? Who is responsible and who is involved?
The answers to these questions enabled decision-makers to take the appropriate actions to
increase safety. The selection of leading metrics (e.g., number of incidents/accidents, or criti-
cal alarms) are also important to reveal cause-and-effect relationships and alert the decision-
makers to focus their immediate attention on the more problematic airport operational proce-
6. Conclusion and Research Direction
The proposed model perceives business processes from a communication, software-driven
activity perspective. That is, we regard business processes as sets of activities constrained,
among others, by causal or spatio-temporal relationships as well as specific business rules. In
particular, the proposed model captures the relationships between different indicators has a
primary role to ensure that effective and reliable information is delivered. The emphasis is put
on the argument that space and time should be considered conceptually equivalent and treated
as such from the modeling point of view.
Adding temporality into spatial data results in the definition of spatial evolutions and helps
in developing and calibrating forecasting models. It also helps in defining the spatio-temporal
resolution for the specific planning problem. Yet, time and space are merely parameters in
which change is possible .
However, the key to conceptual modeling of business dynamic systems is not time nor
space, but change and related constructs such as events and processes. Our proposition is to
exploit and analyze critical business information using the geospatial technology as a data
integrator enabler within enterprise information systems on an ongoing basis.
On the other hand decision-makers usually do not have time and skills to run complex spa-
tio-temporal data queries. Hence, information has to be provided in the form of reports and
dashboards carrying the relevant indicators, as well as through automated alerts activated by
business rules. Such characteristic has leaded us to consider multicriteria decision making,
such as the AHP and the BSC methodology to decompose the decision problem into a hierar-
chy of goals, objectives and criteria.
In our model the business processes are analyzed at different levels from a high view down
to one showing each individual task. The model has to be improved to get a deeper focus on
the representation and propagation of continuous changes at different levels of abstraction,
with an emphasis put on the argument that space and time should be considered conceptually
equivalent and treated as such from the modeling point of view.
Moreover, when analyzing spatio-temporal data many parameters may be considered in the
geoprocessing workflow, making it very difficult for decision-makers to keep track of all the
facts, assumption, and datasets of existing values. Hence, the challenge is now to accomplish
all those requirements in a flexible and efficient way.
This is particularly important because for the decision process more than identifying when
an inconformity occurred, it is interested in understanding the cause of the inconformity. The
challenge resides in providing accurate insight that effectively lead the analyze process to the
identification of the source(s) which most contributed to the observed phenomena. These
sources might drive from internal or external events with some influence in business perform-
Pag. 11 - 12
AIRNET project founded the research detailed in this paper. We gratefully acknowledge the
European Community’s ongoing effort to promote projects which contribute for geospatial
 Abran, A., and Buglione, L. A Multidimensional Performance Model for consolidated Scorecards. In Proc.
Of the 3rd International Workshop on Software and Performance, July 24-27, Rome, Italy, 2002.
 Bédard, Y., et al. Modeling Multiple Representations Into Spatial Data Warehouses: A UML-Based Ap-
proach. Symposium on Geospatial Theory, Ottawa, 2002.
 Bédard, Y., et al. Modeling Geospatial Databases with Plug-Ins for Visual Languages. In Proc. of ER2004,
LNCS 3289, 17–30, 2004.
 Buglione, L., and Abran, A. Balanced ScoreCards and GQM: what are the differences? FESM-AEMES
Software Conference, 2000.
 Chintalapani, G, et. al., Extending the Utility of Treemaps with Flexible Hierarchy. In Proc. of Int. Conf. on
Information Visualisation, London, July 2004.
 Claramunt, C., and Jiang, B. An integrated representation of spatial and temporal relationships between
evolving regions. Journal of Geograph Systems, 3:411–428, 2001.
 CMM. COGNOS Metrics Manager User Guide, COGNOS Enterprise BI Series, http://support.cognos.com,
 Golfarelli, M., Rizzi, S., and Cella, I. Beyond Data Warehousing: What’s Next in Business Intelligence? In
proc of DOLAP’04, November 12–13, Washington, DC, USA, 2004.
 Jensen, C.S., et al. Multidimensional Data Modeling for Location-Based services. VLDB Journal, 13:1-21,
 Malinowski, E. Representing Spatiality in a Conceptual Multidimensional Model, GIS, 2004.
 Microsoft Technet. Business Intelligence and Data Warehousing in SQL Server 2005, July 15,
 Niven, P.R., Balanced Scorecard Step-By-Step: Maximizing Performance and Management Results, John
Wiley & Sons, Inc. 2002)
 Pestana, G., Silva, M. and Bédard, Y.: Spatial OLAP Modeling: An Overview Base on Spatial Objects
Changing over Time. In Proc. of the ICCC (2005)
 Pfoser, D., Pitoura, E., Tryfona, N. Metadata Modeling in a Global Computing Environment. In Proc. of the
ACM-GIS’02, November 8-9, McLean, Virginia, USA (2002)
 Rinner, C., and Taranu, J. A Geographic Visualization Approach to Multi-criteria Decision making, De-
partment of Geography, University of Toronto, 2005.
 Sonnen, D., GVM: A Framework for Estimating the Business Value of Geospatial Technology within In-
formation Systems. Integrated Spatial Solutions, Inc. 2001.
 Vangenot, C. Multi-representation in spatial databases using the MADS conceptual model. ICA Workshop
on Generalisation and Multiple representation, Leicester, 2004.
 Vinturella, B., and Patti, L. Business Process Management (BPM): Must Have?, University of New Orleans, Formatted: Font: 9 pt, Font
April 30, http://mb.jbv.com, 2005. color: Auto, English (U.K.)
 Worboys, M. F., Event-oriented approaches to geographic phenomena, International Journal of Geographi-
Formatted: Font: 9 pt, Font
cal Information Science, 2005.
color: Auto, English (U.K.)
 Yang, Y., A Process-oriented Multi-representation of Gradual changes Journal of Geographic Information
and Decision Analysis, vol. 7, n. 1 pp. 1-13, 2003.
Pag. 12 - 12