Tarek Sboui

1
A Conceptual Framework to Support Semantic
Interoperability of Geospatial Datacubes

• Tarek Sboui
• Yvan Bédard
• Jean Brodeur
• Thierry Badard
• Presented by Yvan Bédard

2
Presentation outline

• Introduction
• Challenges
• Proposed approach
• Semantic heterogeneity conflicts in geospatial
  datacubes
• Framework to support semantic interoperability
  between geospatial datacubes
• Semantic Model
• Conclusion

3
Introduction needs to satisfy
OLAP server
transactional BD
Metadata
Datacube
Used by
Extract Transform Load
Used by
Data Warehouses
Datacubes
Datacube
other sources
Data Marts
Data Sources
DBMS
Clients (Front-End Tools)
OLAP
4
Introduction needs to satisfy

• N heterogeneous cubes must look as only 1
  homogeneous cube
• N heterogeneous cubes must facilitate feeding one
  homogeneous cube

Application, System, User
A simultaneous and rapid navigation through
different datacubes
Insertion of data in a datacube
Interoperability
5
Challenge combine 3 bodies of knowledge
Interoperability

• Promotes data reuse ? gain time and money
• Deals with standards, metadata, ontologies, etc.

Interoperability of geospatial datacubes
Decision Support Systems (DSS)
Datacube
GIS

• Enhance decision making process.
• Detailed data vs. Aggregated data
• Multiples data sources

• Geospatial data (thematic, spatial and temporal)
• Transaction oriented. Not developed for
  strategic decision making.

Several Datacubes
6
Geospatial Datacubes

• Both dimensions and measures may contain
  geospatial components.

Example
Geometric spatial dimension
Mixed spatial dimension
Non-geometric spatial dimension
Canada
Canada

CB

Québec
NB

Montréal
Québec



7
Challenge Newells 10 second cognitive band

• Supports Spatial On-Line Analytical Processing
  (SOLAP)
• Operations such as drill-down, drill-up and
  drill-across.

Select regions -gt Roll-up levels-gt drill-down
4 clicks, 2 seconds
Select 1 year -gt Select all years -gt Select 4
years -gt Multimap View 7 clicks, 5 seconds
Drill down level -gt Change measure -gt roll-up -gt
Roll-up -gt Pivot 8 click, 6 seconds
8
Challenges

• - The content of geospatial datacubes are
  usually heterogeneous
• Organisational heterogeneities
• technical heterogeneities
• Semantic heterogeneities
• - Todays interoperability concepts and standards
  are for transactional systems (they do not
  support multidimensional concepts)

- No work on semantic interoperability of
geospatial datacubes !!
9
Proposed approach

• Defining the semantic interoperability between
  geospatial datacubes
• Categorizing the semantic conflicts that may
  occur between the content of geospatial
  datacubes.
• Defining a framework for interoperating
  geospatial datacubes (based on human
  communication, ontology and context)
• Developing a semantic model that can be used by
  human and machine to
• explicitly represent the elements of datacubes
  semantics.
• interpret datacubes content and reason about
  their semantics.
• Developing a method to reason about the semantics
  of datacubes.
• Developing a method to convert some elements of
  these semantics from source datacube to target
  datacube.

10
Defining the Interoperability of Geospatial
Datacubes

• The interoperability among two geospatial
  datacubes C1 and C2 is the ability of C1 to
  request a service in a manner that can be
  understood by C2, and the ability of C2 to
  respond to that request in a manner that can be
  understood by C1. The request and response are
  conducted automatically.
• Services could include
• importing/exporting data contained in dimensions
  or facts
• Ex. combining forest data from C1 and population
  density from C2 for a risk analysis
• getting information about the dimensions or the
  facts (e.g. language used)
• Ex. Canada national C1 getting data from
  provincial C2s using English or French data
• comparing a dimension/fact against other
  dimension/fact
• Ex. 2004 census C1 with 1999 census C2
• taking into account a concept evolution (e.g.
  meaning or format changes) or
• Ex. 2004 forest inventory C1 combined with 1984
  forest inventory C2 (practices have changed for
  ecological reasons)
• adapting the meaning of a concept when the
  context changes.
• Ex. Fine-grained land use categories in C1 with
  coarse-grained land-use categories in C2

11
Semantic Interoperability of Datacubes VS
Semantic Interoperability of transactional DBs
Semantic interoperability of datacubes
Semantic interoperability of transactional DBs
Similarities
Reusing data Facilitates an efficient exchange of
information
Differences
Deals with datacubes concepts (facts, measures,
dimensions, levels) Deals with the semantic
heterogeneities of of aggregation and
summarizing methods and algorithms, including
summarizability conditions.
Deals with the heterogeneities of DB concepts
(i.e. tables, attributes, relations, etc.).
12
Categorizing the semantic heterogeneities
13
Human Communication, Ontology and Context
Field of experience
Field of experience
Signal
Encoder
Decoder
Source
Destination

• An ontology is a set of related concepts and a
  set of assumptions about the intended meaning of
  these concepts in a given domain or application.
• Context is any information that surrounds and
  facilitates the interpretation of concepts.
• Ontologies contain only certain elements of
  context.

14
Human Communication, Ontology and Context

• We identify four context levels
• Concept Context level includes the
  characteristics of dimensions or measures of
  datacubes (role, properties, etc).
• Dataset Context level consists of elements
  related to dataset of geospatial datacubes (such
  as the specifications used to describe concepts),
  
• Domain Context level contains the context
  elements of the domain (such as forestry),
• Goal Context level defines the purpose for which
  the geospatial datacubes will be used (such as
  the evolution of the wood volume),

15
Interoperability Framework
Inferred contexts
Context Agent
Agent B (Datacube B)
Agent A (Datacube A)
Context ontology
Context ontology
Datacube A ontology
Datacube B ontology
Datacube A
Metadata A
Metadata B
Datacube B
16
SemEL A Semantic Model for Interoperating
Geospatial Datacubes
Definition
Global Context
GD2
House Construction for living
Spatial Context
DD11
DD12
DD21
DD22
red
Colour
UML
Domain Context
Application Domain
Data Source Context
Technique
Graphic
blue
C1
C2
O1
O2
Year
Time
Month
E/R
C
1K
O
1K
1/5000
scale
Graph
Description
1/1250
NL
Assertion
Assertion
Constraints
Prosperities
Polygon
line
point
Context of concepts
Geometry
17
Conclusion

• Our approach is based on human communication,
  ontology, context, and the multidimensional
  structure.
• We defined a communication model which is based
  on Datacubes Agents and a Context Agent.
• The SemEL model represents a new concept that
• is based on the multidimensional paradigm.
• explicitly represents the semantics of geospatial
  datacubes contents.
• allows to reason about semantics.
• can be implemented on a relational platform.

18
Current Future Works

• Current works
• SemEL Refined
• Definining a method to reason about the semantics
  of geospatial datacubes
• Based on SemEL
• Require human intervention
• Future works
• Define a mapping method between the content of
  the geospatial datacubes contents.
• Develop a SOLAP application based on SemEL.
• Develop a web-based system that supports the
  mapping between datacubes contents.

19

• Thanks !

20
Geospatial Datacubes
Datacubes Roads and Lakes
Dimension Road
Level City
Level Province
270

180
240
Montreal
Dimension Region
Fact relation between roads and lakes
Quebec
Quebec
Int.
Adj.
Adj.
Disj.
Toronto
Adj.
Adj.
Int.
Adj.
Disj.
Ontario
Disj.
Adj.
Int.
Disj.
Ottawa
2003
Adj.
Adj.
Disj.
In.
2002
Dimension Time
2001
Lac Beauport
Lac St-Charles
Members of dimension
Dimension Lake
The raison dêtre of datacubes ? Support
strategic decision making