The Rise of Model-Based Data and Its Implications for Social Science and Policy

Explore the fundamental shift in social science data and its implications for urban planning, research, and policy-making in this keynote address. Delve into the concept of "modelification" of data, examining how model outputs are increasingly replacing direct observations in social sciences. Understand the factors accelerating this transition, including differential privacy concerns, politicization of national statistical systems, and survey non-response. Discover the impact of this epistemic shift on urban analytics, planning practices, and social scientific inference. Learn about the challenges and opportunities presented by model-based data in the context of urban development and policy formulation. Gain insights into the future of urban analytics and its intersection with artificial intelligence. Engage with thought-provoking examples from the American Community Survey, the 2020 census, and COVID-19 income differences to illustrate the complexities of model-based data.

Introduction
Overview
Definitions
American Community Survey
Variation in the model
Uncertainty
The 2020 census
Accuracy vs privacy
Topdown algorithm
Liberty Island
Covid19 and income differences
Experiments
Wrapup
What does this mean for urban analytics
What does this mean for AI
Questions
Survey Responses
Discussion
Privilege of the boss
Uncertainty in data