Predicting neighborhood change using big data and machine learning: Implications for theory, methods, and practice

Urban Displacement Project


January 9, 2020 to January 10, 2020
9:15am to 4:30pm
UC Berkeley

This symposium will convene an international group of urban researchers with deep interests in data science and neighborhood change. Despite decades of research on neighborhood change, there has been little corresponding methodological development: studies still tend to either rely primarily on demographic data aggregated at the neighborhood level (which masks complex and micro-scale causal dynamics), or on in-depth case studies (which present challenges for generalization). Advances in data science, particularly if informed by critical urban theory, offer the potential to remedy some of these methodological shortcomings. For instance, real-time data on activity patterns, such as geotagged tweets, can help overturn traditional conceptions of residential segregation (Shelton, Poorthuis, and Zook 2015), and bridge time lags in census data (Hristova et al., 2016). Using machine learning techniques, we can also analyze existing patterns of neighborhood ascent and decline in order to predict future change (Reades, de Souza, and Hubbard, 2019). To the extent that these and other approaches support an early warning system designed to be readily understood by stakeholders, they have the ability to empower communities, at a minimum, and potentially to transform policy as well (Chapple and Zuk 2016).

Symposium events will held at UC Berkeley on January 9-10, 2020, and at the University of Sydney on June 1-2, 2020, with each two-day program consisting of a mix of keynote speakers, seminars, panels, and workshops with data science researchers and government officials. We expect to publish the results of our work in a special issue of a peer-reviewed journal, to be determined.

This conference is full and registration is closed.  Please contact Karen Chapple ( for more information about this initiative and/or to request to be added to the mailing list for the Sydney meeting.

CALL FOR ABSTRACTS: Submissions due by September 20, 2019
We are seeking papers about neighborhood change that innovate by using user-generated geographic information, social media data, machine learning, image processing, or the like. We are particularly interested in theoretically informed studies that adopt a comparative lens or mixed methods.  If you would like to present a paper, please submit an abstract of no more than 500 words by September 20, 2019 to Please specify which conference you would like to attend: Berkeley, Sydney, or both. Unfortunately we cannot offer any funding to support travel. Authors of the selected abstracts will be notified by early October and be expected to submit their completed papers by one week before each conference.

Conference Organizers
Project Leads
Karen Chapple; Professor and Chair, City and Regional Planning; University of California, Berkeley
Nicole Gurran, Professor and Chair, Urban and Regional Planning and Policy; University of Sydney
Somwrita Sarkar; Senior Lecturer, Design; University of Sydney
Project Team
Cynthia Goytia; Professor and Director, Urban Economics; Universidad Torcuato di Tella
Ate Poorthuis; Assistant Professor, Geography; Singapore University of Technology and Design
Jon Reades; Senior Lecturer, Quantitative Human Geography; King’s College, London
Matthew Zook; Professor and Interim Chair, Geography; University of Kentucky

Karen Chapple; Professor and Chair, City and Regional Planning; UC Berkeley -

This conference is made possible with support from the Urban Studies Foundation, the University of California-Berkeley, and the University of Sydney. At Berkeley, the conference will be held at the Berkeley Institute of Data Science, which is a co-sponsor of the Berkeley event. The Urban Displacement Project at UC Berkeley aims to understand the nature of gentrification and displacement in American cities, focusing on creating tools to help communities identify the pressures surrounding them and take more effective action