In the fall semester 2015, I attended the lecture “geographic information retrieval”. As part of the assessment process, a paper had to be written. I wrote a paper on a conceptual model for a graph driven gazetteer.
In this paper, I describe the state of the art in geographic information retrieval (GIR) and present a conceptual model for a fast, scalable graph based gazetteer, to be used in various domains of geographic information retrieval. The majority of available gazetteers are stored as relational databases, providing easy access but limiting the complexity of queries. In this paper I create and discuss a conceptual model of a gazetteer using a Neo4J graph database, which allows queries of high complexity to be efficiently run with low response time. A Neo4J graph database is installed onto a Linux server and is populated with a limited dataset to illustrate different aspects of the conceptual model. Queries are written using the generic Neo4J querying language CYPHER. Special focus is given to storing and querying semantic relationships between locations to allow queries about vague spatial relations on different scales. The proposed conceptual model has the potential to solve various state of the art limitations of available gazetteers discussed in this paper, but is accompanied by new limitations needing further attention.