Open Data and machine learning in the service of complementing municipal GIS inventory
Keywords: Urban GIS, machine learning, open data
Abstract. In this study the authors investigated the possibilities to use open data and open software complemented with machine learning to enhance the content of municipal databases. In the study area in Székesfehérvár, a GIS system is used with approximately with 30 modules, although many are still missing. The authors prepared examine the easiest and most affordable methods to extract data to use in two future modules: Parking and Traffic Engineering module. In parking model along field survey, they used QGIS and OpenStreetMap, in the other module they used Google StreetView for defining the places of traffic signs and used machine learning to define the signposts. They found that the accuracy of creating the parking module is based on the completeness of the database and the field measurement method, in case of the Traffic Engineering method the up-to-dateness and completeness of the original data source (Google Street View) and the number of teaching samples influence the results.