Statistics analysis embedded in spatial DBMS

Abstract

This article sets forth the principle and methodology for implementing spatial database management system (DBMS) by using open source object-relational DBMS - PostgreSQL. The geospatial data model and spatial analysis and processing operations for spatial objects and datasets can be inserted into the DBMS by extended SQL. To implement the statistics analysis embedded in spatial DBMS, an open source statistical program R is introduced to extend the capability of the spatial DBMS. R is a language and environment for statistical computing and graphics. There is a large sum of statistical methods in the form of packages in R. Many classical and modern spatial statistical techniques are implemented in R environment. PL/R is a loadable procedural language containing most of the capabilities in R language which is extensible and enables user to write DBMS functions and triggers in R language. Therefore, the PL/R will extend its capability of spatial statistics and geostatistics when the two kinds of packages are loaded into R language. The PL/R can be extended without limit so that any new method of statistics analysis embedded into the spatial DBMS becomes very convenient.