Individual accessibility and segregation on activity spaces: an agent-based modelling approach

Pereira Saraiva, Marcus Vinicius (2020) Individual accessibility and segregation on activity spaces: an agent-based modelling approach. Doctoral thesis, Birkbeck, University of London.

Abstract

One of the main challenges of cities is the increasing social inequality imposed by the way population groups, jobs, amenities and services, as well as the transportation infrastructure, are distributed across urban space. In this thesis, the concepts of accessibility and segregation are used to study these inequalities. They can be defined as the interaction of individuals with urban opportunities and with individuals from other population groups, respectively. Interactions are made possible by people’s activities and movement within a city, which characterise accessibility and segregation as inherently dynamic and individual-based concepts. Nevertheless, they are largely studied from a static and place-based perspective. This thesis proposes an analytical and exploratory framework for studying individual-based accessibility and segregation in cities using individuals’ travel trajectories in space and time. An agent-based simulation model was developed to generate individual trajectories dynamically, employing standard datasets such as census and OD matrices and allowing for multiple perspectives of analysis by grouping individuals based on their attributes. The model’s ability to simulate people’s trajectories realistically was validated through systematic sensitivity tests and statistical comparison with real-world trajectories from Rio de Janeiro, Brazil, and travel times from London, UK. The approach was applied to two exploratory studies: São Paulo, Brazil, and London, UK. The first revealed inequalities in accessibility by income, education and gender and also unveiled within-group differences beyond place-based patterns. The latter explored ethnic segregation, unveiling patterns of potential interaction among ethnic groups in the urban space beyond their residential and workplace locations. Those studies demonstrated how inequality in accessibility and segregation can be studied both at large metropolitan scales and at fine level of detail, using standard datasets, with modest computational requirements and ease of operationalisation. The proposed approach opens up avenues for the study of complex dynamics of interaction of urban populations in a variety of urban contexts.

Metadata

Statistics

DownloadsShow export options

Export as XMLJSONCSV

Activity Overview

6 month trend

120Downloads

6 month trend

148Hits

Additional statistics are available via IRStats2.

Archive Staff Only (login required)

Edit/View Item