A New Picture of the City: Volunteered Geographic Image Information and the Cities



Chen, Meixu ORCID: 0000-0003-2712-5551
(2021) A New Picture of the City: Volunteered Geographic Image Information and the Cities. PhD thesis, University of Liverpool.

[img] Text
201228789_Mar2021.pdf - Unspecified

Download (4MB) | Preview

Abstract

The urbanisation process continuously influences human life, causing long-term challenges for the planning and management of urban areas. In recent years, with the emergence of new forms of data and advances in techniques, the ways of managing and governing this process have evolved and formed a new research field: urban analytics. A growing number of human behaviours can be traced through quantities of data, which enables attributes of the urban environment to be managed more efficiently, potentially beneficial to complex decision-making processes by stakeholders. As such, how to extract useful information from new data and provide more suitable methods requires careful consideration. The question of how human activity relates to the built environment has been an important topic in the sensing of cities. Existing ways to perceive the city either focus on environmental aspects that cover historical, social, or cultural dimensions of urban space through surveys, interviews, or mobility data (e.g., social media data), or extract visible features from georeferenced images to gain perceptions of the city. However, both approaches are often disconnected and lack dynamic consideration. The main aim of this thesis is to address these challenges and gaps within urban analytics. It develops a methodological framework to leverage user-generated geotagged images and modern analytical techniques to obtain insights. Such framework is designed to mine spatial, temporal and image attributes of the Flickr images, which combines multiple dimensions including spatiotemporal dynamic analysis, computer vision models, summary statistics, and varying machine learning algorithms that allow understanding of human interactions with the built environment. The overall analysis and results enrich our current understanding of how user-generated urban pictures represent but also shape the city. This is especially important given the growing popularity of volunteered geographic information and urban analytics over the last decade. Their rapid growth has facilitated debates worldwide, but there is still a large potential of volunteered geographic information such as geotagged image information which has been underestimated in most circumstances. The findings presented in this thesis offer richer evidence that aims to help the improvement of strategic planning systems, and empowering policymakers to make smarter decisions in terms of urban governance.

Item Type: Thesis (PhD)
Uncontrolled Keywords: volunteered geographic information, geotagged Flickr images, urban perception, urban areas of interest, machine learning
Divisions: Faculty of Science and Engineering > School of Environmental Sciences
Depositing User: Symplectic Admin
Date Deposited: 13 Jan 2022 14:29
Last Modified: 18 Jan 2023 21:24
DOI: 10.17638/03143011
Supervisors:
URI: https://livrepository.liverpool.ac.uk/id/eprint/3143011