Understanding Semantics and Ontologies in Spatial Database Applications

To understand the concept of semantics in spatial database applications picture this scenario: Joan has been tasked with mapping points of interest in her neighborhood.  Before mapping, she creates categories for these points of interests.  She chooses restaurants, schools, clothing stores, gas stations, parks, dry cleaners, etc.  For each of those entities that exist in her neighborhood she drops a point on the map and assigns the appropriate category.  Across town, John is sitting in his office and is working on creating a major point file for the whole county.  John has a large task and creates broader categories like commercial, recreation, industrial, education, etc.  Joan and John both put a point on the High School but Joan categorizes it as a school and John marks it as education.  This scenario provides a very general understanding to the concept of semantics, meaning “there is no right or wrong description, just different meanings for different purposes for different people” (Hunter, 2002, p.85).  Let’s say that Joan and John were tasked to merge their point of interest and major point file to create a searchable database to be published on the web and accessible to everyone.  This would give an opportunity to introduce the concept of ontology.  Instead of breaking down one of their databases to conform to the other, Joan and John “could integrate several ontologies so that [they] in fact create a new formal description of how separate databases compare to each other (without actually changing the nature of the databases) (Hunter, 2002, p.86).  Integrating ontologies would allow them to keep both of their databases as is and provide a seamless searchable database to the end users.

In the example above Joan is a citizen who wants to map out points of interest in her community for informational purposes while John in a mapping professional who needs to map major points in the county he works in for an upcoming project.  Citizens and professionals represent two different types of “information communities”.  A citizen of the community, visitor to the area, or a potential citizen wants to see places they can eat supper tonight or go shopping.  Planners, Realtors, Administrators and other professionals may only need to identify and analyze commercial corridors in the area.  Every information community is looking at the same spatial entity but interpreting and analyzing it in very different ways.  When performing a search a hungry student may enter restaurants while a city planner may enter commercial buildings and they would both get the same results.  This reinforces the concept that “ontologies can be used to establish agreements about diverse views of the world and consequently carry the meaning of the original ideas that are embedded in the representation of geographic phenomena in the human mind” (Fonseca, 2002, p.146).  

“To adequately represent the geographic world, we must have computer representations capable of not only capturing descriptive attributes about its concepts, but also capable of describing the geometrical and positional components of these concepts” (Fonseca, 2003, p.356). Many times the nature in which points of interests in a community are used determines how they should be represented on a map.  A realtor or potential citizen may only need a single point to determine how many schools are located within a mile of a house.  A student or visitor to the school would like to see the building footprints on the school campus to determine where they would need to go.  The school district administrator would like to see the campus boundaries in order to determine what the acreage is.  These three examples are of the same spatial entity (a school) but represented and understood in three very different ways.  Identifying and understanding semantics is a good place to start in order to determine and integrate ontologies into a spatial database.  This is easier said than done, but as our world becomes more connected we find the need to do this more important.    

References

Hunter, G.J. (2002, March).  Understanding Semantics and Ontologies: They’re Quite Simple Really – If You Know What I Mean!  Transactions in GIS 6(2), 83-87.

doi: http://dx.doi.org/10.1111/1467-9671.0009

Foncesca, Frederico, Davis, Clodoveu, & Câmara, Gilbert (2003, December).  Bridging Ontologies and Conceptual Schemas in Geographic Information Integration.  Geoinformatica 7(4), 355-378.

Retrieved from http://link.springer.com.libproxy.usc.edu/article/10.1023%2FA%3A1025573406389

Foncesca, Frederico, Egenhofer, Max, Davis, Clodoveu, & Câmara, Gilberto (2002, September).  Semantic Granularity in Ontology-Driven Geographic Information Systems.   Annals of Mathematics and Artificial Intelligence 36(1-2), 121-151.  Retrieved from

               http://link.springer.com.libproxy.usc.edu/article/10.1023%2FA%3A1015808104769