Determining the data quality-influencing characteristics of participants in volunteered geographic information campaigns – the case of maritime disasters

This study analyses how participant characteristics influence the quality of participants in volunteered geographic information (VGI). Despite the wide use of VGI in different areas, the quality of the data collected can vary significantly due to the diversity of participants. In the study presented here, we used data collected as part of the VGI campaign to simulate maritime disasters to investigate how participant characteristics affect the quality of VGI. We use machine learning, specifically the Random Forest algorithm, to analyse the impact of factors such as demographics, technical expertise and geographical knowledge on the VGI quality metrics defined by ISO 19157 (positional accuracy, completeness and thematic accuracy). This study aims to: (1) identify key participant characteristics that impact VGI quality, (2) evaluate the effectiveness of Random Forest in predicting VGI quality based on these characteristics, and (3) provide recommendations for improving VGI campaigns through participant selection, training, and data collection methods. By understanding the influence of participants, we aim to realise the full potential of VGI and provide reliable and high-quality data for scientific research and practical applications, for the maritime domains.