Programme de recherche

Voici ce qui dĂ©crivait mon programme de recherche des dernières annĂ©es oĂą j’Ă©tais Ă  l’UniversitĂ©

Research Program

My research program aims at improving the usability of geospatial databases through better design, empowering users with geospatial business intelligence technology, and managing the risks of data misuses. Fundamental research is typically tested and iteratively improved over the years with real-life projects in industry via contracts or cooperation grants from Canada NSERC, FCI, GEOIDE, etc. A multidisciplinary approach allows us to test our concepts in several fields to ultimately lead to fundamental concepts that are generic and reusable.

Improving the usability of geospatial databases through better design

In this area, research has focused on extending the Unified Modeling Language (UML) to reduce the size and complexity of geospatial database conceptual models while improving their readibility and robustness. Proposed solutions have been tested internationally in hundreds for several years, both internally or by outside organizations and private consultants, providing us with feedbacks and suggestions that led to iterative improvements. Today’s solution is presented in a new web site and has taken the form of a pictogrammic language called PictrograF. PictograF can be used with several categories of tools, methods and standards (ex. computer-assisted software engineering, human interface design, reporting, metadata engine, specifications, ISO19xxx compliant models, OGC-compliant applications, plain text). PictograF facilitates and improves the description of the geometry and temporality of any feature, of any attribute and of any relationship to be found in a 2D or 3D geospatial database or document. To have a description of PictograF and examples of its usages in different situations and CASE technologies, please visit URL de PictograF. We also developed a UML-based CASE tool for high-level conceptual database models, it is called Perceptory (mettre URL du nouveau site) and is meant primarily to be used to design Computation-Independent Models (CIM) as suggested in the MDA standard (Model-Driven Architecture) of the OMG (Object Management Group). The ancestor of Perceptory was called Orion (1992), it was the world first CASE tool to automatically generate the code for a spatial database structure in commercial GIS technologies (cf. Intergraph TIGRIS, MGE, MapInfo, Oracle Spatial). It was based on the Extended Entity-Relationship language used in the MERISE system design method.

Empowering users with geospatial business intelligence technology

Once the database is designed and implemented, one must look for the best return on investment in these data. While GIS technology and geospatial DBMS do an excellent job at data storing, updating, integrity checking and basic querying, they fall short in comparison to GeoBI technologies (Geospatial Business Intelligence) for complex queries producing summarized information, trends analysis, spatio-temporal comparisons and correlation synthesis, geographic knowledge discovery or similar operations involving fine-grained to highly aggregated information in a multi-scale, multi-epoch and multi-theme analysis. We have been pioneering the research in this field since the mid-1990s after doing research for Canada Defence. I coined the acronym SOLAP in 1996 (Spatial OLAP) to explicitely include spatial capabilities to the emerging OLAP technology (N.B. it was E.F. Codd who coined the term OLAP in 1993 for On-Line Analytical Processing, as opposed to OLTP for On-Line Transaction Processing). Today’s solution is available in the software Map4Decision (http://www.intelli3.com ), formerly JMap SOLAP, and the fundamental concepts are described in several papers available in the « publication » section of my web site. This solution extends OLAP servers with capabilities to interactively explore multi-granularity data in maps as well as in tables and charts. This solution has several unique cartographic capabilities and spatial operators. Furthermore, it was designed to read existing GIS and spatial DBMS data, adding value to these data by allowing their users to easily explore interactively their data, to produce summarized information, to analyse trends, to perform spatio-temporal comparisons, etc. By not requiring to learn a query language, these new capabilities allow a larger number of users to benefit immediately from spatial data since only a few clicks and a few seconds are required to obtain the desired information. This is true users empowering.

Managing the risks of geospatial data misuses

Once data are available to a larger audience, the risks of data misuses increase. This has been a regular concern of some of our users. It is also one of the top priorities of the BI community as well as an increasing concern of the geomatics industry. We have found numerous cases of incidents and accidents resulting from spatial data misuses. We have found several Court decisions and official regulations restricting the use of spatial data or rendering data producers more liable. Now that geospatial data and technologies have been democratized, Society is evolving towards more preventive approaches. As geospatial database designers and technology providers, we have been doing research to better protect the users against potential misuses. Today’s solutions involve a consumer protection awareness and a risk-management approach for system design and data production. This includes the implementation of context-sensitive warnings in applications, the development of user manuals explicitely embedding risk-prevention information, the development of quality audit methods, of methods to assess fitness-for-use, of extended warranties, of user-meaningful metadata, etc. More information is found in the papers published, which are available in my web site, as well as in http://dataquality.scg.ulaval.ca .

Real-life testing

Most concepts developed by the research team have been tested and improved in collaboration with more than 50 organizations since I started as a professor. Since we aim at developing generic concepts that are applicable to several fields of application, we believe than GIS research needs to be tested with several real usages and users to be robust. GIS are designed, implemented and used by humans who have uncertain knowledge about the Reality observed in their field, they work within organizations under resource constraints. Research methods are influenced by these factors. Accordingly, we involve organizations once or several times in our projects such as include Defence Canada, Natural Resources Canada, Health Canada, Infrastructures Canada, Geomatics Canada, Canada National Collaborating Centre for Environmental Health, Quebec Department of Transportation, Natural Resources and Wildlife QuĂ©bec, Quebec Public Health Institute, Quebec Department of Health and Social Services, Quebec Department of Environnement, Quebec Department of Public Security, Hydro-QuĂ©bec, DMR Group, K2 GĂ©ospatial, Intergraph, ViaSat, Syntel, Alta Group, IntĂ©lec GĂ©omatique, Holonics, Harfang, EMCO, SNC-Lavalin, British-Columbia Ministry of Environment, Lands and Parks, North American Wetlands Conservation Council, TSI (France), AECOM, MITRE (US), World Bank, etc. It also helps to do research when we have personal experience outside of academia, accordingly I had the early privilege to work from 1977 to 1980 as an assistant-land surveyor for two private firms and for the Cadastral Service of the Quebec Government. In addition, I’ve been doing occasional consulting in GIS and Spatial database since 1980. Finally, the core members of the research team created a start-up company called Intelli3 where additional experience is continuously acquired.

Multidisciplinary approach

It has always been my objective to test our concepts in very different fields. It is an excellent way to see their completeness or robustness in face if varying situations. Accordingly, while our main fields of expertise have been transportation, topographic databases and public health, we have done projects in agriculture, forestry, environment, wetlands, erosion risks management, industrial risks management, health and climate change, public health, kinesiology for olympic sports, cadastre, property assessment, power lines, geospatial data catalogues and repositories, water and sewer systems, infrastructures project management, security, archeology, ontology management, etc. To do so, we have collaborated with other researchers from Laval University (computing sciences, agriculture, forestry, archeology, land planning, administration, etc.) and from other universities or academic organisations: CEMAGREF (Montpellier and Clermont-Ferrant, engineers and computer scientists : GIS, SOLAP and software engineering in agriculture), INSA Lyon (computer scientists : SOLAP in forestry), IGA (Univ. Grenoble, geographers: GIS in natural risks), Univ. Angers (geographers: GIS and geomatisation), IGN-Paris (engineers and computer scientists: GIS et generalization), ParisTech (engineers : Sophia-Antipolis : SOLAP in natural risks), Univ. Strasbourg (geographers: SOLAP in industrial risks), EPFL Switzerland (land surveyors: GIS, GIS implementation, software engineering and SOLAP), Univ. Marne-La-Vallée (computer scientists: data quality), Univ. Costa-Rica (computer scientists : SOLAP), Univ. New-Brunswick (Geomatics engineers, computer scientists: GIS, SOLAP and data quality), Memorial Univ. (geography: data quality), McGill Univ. (computer scientists: SOLAP), Simon-Fraser Univ. (computer scientists: SOLAP), Univ. Calgary (Geomatics engineers : SOLAP, GIS), Univ. Ottawa (lawyers: data quality), Univ. Sherbrooke : (doctors: SOLAP in public health), Univ. Genève (geography: SOLAP and thematic cartography), Univ. Montpellier (computer scientists: GIS), Univ. Lyon II (computer scientists: SOLAP), Univ. McMaster (geographers: public health), etc.

Insofar, this program and these experiments have allowed my research team and I to deliver over 150 refereed publications and 225 conferences worldwide, advanced graduate courses, over 200 graduates, continuing education for 1500 professionnels on 3 continents, transfers of technologies and methodologies in several organizations, one commercial product (Map4Decision) and our start-up company (intelli3.com).