UGOD Seminar | Garbage In, Gospel Out? Towards a better understanding of urban datasets
Supporting the below United Nations Sustainable Development Goals:支持以下聯合國可持續發展目標：支持以下联合国可持续发展目标：
While the quality and availability of predictive modelling methods for urban analyses continue to improve rapidly, our collective understanding of their underlying data remains poor. This lack of insight hampers the practical implementation of methods and reduces our ability to extract knowledge from urban research efforts. Key problems of datasets for predictive modelling include a lack of insight into their veracity, a lack of open-source alternatives, and an insistence on using established dataset paradigms over inventing new ones. In this talk Dr. Alex Levering will present and discuss research efforts focused on improving our understanding of data underpinning urban analyses, as well as research efforts that attempt to establish new dataset paradigms. In particular, the talk will cover research efforts on urban liveability, landscape aesthetic preferences, and urban built-up land datasets.
Dr. Alex Levering is an assistant professor of environmental geography working at the Institute for Environmental Sciences (IVM) of the Vrije Universiteit Amsterdam (VUA). He is also the coordinator of the UNIGIS Amsterdam programme, a GIS-focused postgraduate masters for working professionals. In 2024 he obtained his PhD degree from Wageningen University on the topic of deep learning for landscape qualities. His research is focused on modelling, mapping, and studying socio-cultural landscape interactions. He specializes in landscape qualities such as urban liveability and landscape aesthetic beauty, especially as they are perceived by people, and as measured with the help of deep learning methods. He is especially interested in the fairness, scalability, multiplicity, and variety of datasets underpinning landscape analyses, especially in urban settings.