THE MAKING OF STATISTICAL MAPS: A PERSPECTIVE FROM THE ONTOLOGY OF POLITICAL DISTRICTS

摘要

Statistical information collected by professional domains serves as a valuable reference to understand our world. Often stored in table formats, such data lacks a direct geospatial reference to make it illustratable on a map. In the past, such a geo-linking process is often completed by geospatial professionals via common geographic identifiers between the two datasets. Without such an in-depth knowledge and skills, inexperienced users may find it difficult to correctly pinpoint their data on the maps. From an ontology perspective, we argued that these required geo-links can be automatically established as long as the knowledge about political district and its links with users' statistical data can be "correctly" formalized. Following the concept of ontology, we first analyze how political district data is generated, managed and updated, and then formalize our findings for the design of political district ontology. The ontology enables us to unambiguously describe each political districts and their hierarchies or neighboring relations in the OWL format. The proposed ontology should be established and maintained by the authorized organizations to ensure its unique identification and correctness. The statistical map making mechanism has built-in knowledge to intelligently determine if the generated data can be used for mapping or even being used for producing other types of statistical maps. This collaborative framework is developed upon a number of services respectively managed by different organizations, which not only demonstrate its potential in the future SOA environment, but also perfectly reflect a better division of works for the current governments.