Show simple item record

A workflow for geocoding South African addresses

dc.contributor.advisorBerman, Soniaen_ZA
dc.contributor.authorVan Rensburg, Alexandriaen_ZA
dc.date.accessioned2016-01-02T05:21:50Z
dc.date.accessioned2018-11-26T13:53:57Z
dc.date.available2016-01-02T05:21:50Z
dc.date.available2018-11-26T13:53:57Z
dc.date.issued2015en_ZA
dc.identifier.urihttp://hdl.handle.net/11427/16198
dc.identifier.urihttp://repository.aust.edu.ng/xmlui/handle/11427/16198
dc.description.abstractThere are many industries that have long been utilizing Geographical Information Systems (GIS) for spatial analysis. In many parts of the world, it has gained less popularity because of inaccurate geocoding methods and a lack of data standardization. Commercial services can also be expensive and as such, smaller businesses have been reluctant to make a financial commitment to spatial analytics. This thesis discusses the challenges specific to South Africa as well as the challenges inherent in bad address data. The main goal of this research is to highlight the potential error rates of geocoded user-captured address data and to provide a workflow that can be followed to reduce the error rate without intensive manual data cleansing. We developed a six step workflow and software package to prepare address data for spatial analysis and determine the potential error rate. We used three methods of geocoding: a gazetteer postal code file, a free web API and an international commercial product. To protect the privacy of the clients and the businesses, addresses were aggregated with precision to a postcode or suburb centroid. Geocoding results were analysed before and after each step. Two businesses were analysed, a mid-large scale business with a large structured client address database and a small private business with a 20 year old unstructured client address database. The companies are from two completely different industries, the larger being in the financial industry and the smaller company an independent magazine in publishing.en_ZA
dc.language.isoengen_ZA
dc.subject.otherComputer Scienceen_ZA
dc.titleA workflow for geocoding South African addressesen_ZA
dc.typeThesisen_ZA
dc.type.qualificationlevelMastersen_ZA
dc.type.qualificationnameMPhilen_ZA
dc.publisher.institutionUniversity of Cape Town
dc.publisher.facultyFaculty of Scienceen_ZA
dc.publisher.departmentDepartment of Computer Scienceen_ZA


Files in this item

FilesSizeFormatView
thesis_sci_2015_van_rensburg_alexandria.pdf2.841Mbapplication/pdfView/Open

This item appears in the following Collection(s)

Show simple item record