Mining spatiotemporal growth pattern of volunteered data using a contributor-based approach

Abstract

Volunteered geographic information (VGI) is a valuable data source that can be applied in many fields. Monitoring its spatiotemporal changes is an important issue. Given the unique role of participants, their behaviours significantly influence the growth conditions of the volunteered data. This study proposes a new contributor-based approach to explore the growth pattern of VGI through a three-step model. The first step defines a set of properties and categorizes the contributors. In the second step, the study region is divided into regular cells, and a set of indicators is defined to describe the contribution status of each cell and its neighbouring cells. In the third step, an artificial neural network-cellular automata model is used to extract the pattern of growth. To evaluate the proposed method, it was applied to OpenStreetMap data of a central region of Tehran. According to the confusion matrix, the overall accuracy of 85.3% was achieved with a Kappa coefficient of 0.77.

Keywords:

• Volunteered geographic information
• OpenStreetMap
• growth pattern
• contributor
• ANN-CA

Acknowledgement

This work is published as part of a research project supported by the Iran National Science Foundation (INSF).

Disclosure statement

No potential conflict of interest was reported by the author(s).