Toward correctness control of postal addresses geocoding

The paper discusses content and results of the methodology elaborated for geocoding of postal addresses. The geocoding issue is considered on the example of study devoted to the exploration of the spatial distribution and dynamics tuberculosis and concomitant diseases infection cases. The study is carried out on a large city scale. The example of St. Petersburg city (Russia) is used. Proposed methodology is based upon the extending of the classical geocoding scheme that assumes direct linking of the address data presented as a part of initial dataset with the address data presented in the reference dataset (in the geospatial database). The extension consists in the use of a middle reference-standard register of postal addresses. An address database developed by official agencies is used as a reference-standard register. Initial data records are linked with the records of the reference-standard register by postal addresses, and the register records, in turn, are linked with the records of the reference dataset used for geocoding (with the addresses recorded in the attributes of the geospatial database objects). This approach allows to provide control of structure and content correctness for the address data used for geocoding, as well as to convert address data in to a unified form accepted officially at the state level and used in official documents and information systems. The methodology is implemented on the example of the postal address system used in the Russia. However, it can also be used when operating with the address system of any other states. In such a case, it is necessary to have an official register of postal addresses to implement the methodology. The register have to be presented in a structured form (preferably in the form of a database). The methodology can be used both for medical statistics data geocoding, and for geocoding of other domain data.