Intrinsic Parameters based Quality Assessment of Indian OpenStreetMap Dataset using Supervised Learning Technique

Accuracy of the data plays a crucial role in the effective working of data-driven systems. OpenStreetMap being the source of a spatial database for many location-based services highly contributes towards their performance. OpenStreetMap is a volunteered, non-proprietary dataset so it is more vulnerable to errors and discrepancies. To use the OpenStreetMap data for location-based services, it is mandatory that data should not suffer from topological and geometrical errors. In this paper, topological errors associated with different objects in OpenStreetMap (OSM) data are detected. OSM data of Punjab and Haryana (India) has been taken as test data for finding topological errors. This study is focused on developing a framework for augmenting the topological consistency of OSM data by users. A supervised decision tree approach is presented to find the topological errors in the OSM database. The framework uses REST APIs for communication of data to and from the OSM server. The outcome of this study would certainly help the users to improve the quality of OSM data.