Regions and places in the city: gathering insights from primary and secondary data sources in light of GIScience

Abstract

This thesis investigates the integration of place-based concepts within Geographic Information Science (GIScience) through mixed-methods approaches that leverage user-generated content (UGC) and participatory mapping. The primary challenge addressed is the operationalization of place-based knowledge within Geographic Information Systems (GIS), particularly in urban environments where traditional data sources often fail to capture the nuanced dimensions of places. Despite advancements in GIScience, significant gaps remain in methods that integrate spatial and semantic aspects of places. Existing research often simplifies place-based constructs, such as neighborhoods, neglecting important facets related to human experiences and perceptions. To address these gaps, this thesis presents a series of studies focused on specific research questions related to the operationalization of place within GIS. The research harnesses the potential of both geographic UGC and participatory mapping approaches as essential data sources and are grounded in case studies conducted in the city of Lisbon (Portugal). By developing a framework based on transformer-based topic modeling and spatial statistics, we were able to unravel the relationships between social media content and place-based semantics, providing important insights into content-location relationships. Then, we leveraged multiple UGC sources to extract and compare neighborhood boundaries, highlighting the challenges and limitations of relying solely on UGC for retrieving spatial projections of place-based concepts. Lastly, we employ participatory mapping to explore the richness of residents’ perceptions of neighborhood boundaries as well as their relationships with features of the urban form, particularly street networks. The findings of this thesis contribute significantly to GIScience by providing innovative methodologies for analyzing and mapping place-based concepts. The methods and implementations enhance the understanding of how residents mentally structure their city both spatially and semantically, ultimately supporting more informed urban planning and policy-making. In addition, the thesis critically evaluates the strengths and limitations of both UGC and participatory mapping, underscoring the potential for more accurate and context-sensitive GIS models that reflect the lived experiences of residents.