Outlooks and Challenges in Urban Design X Data Science Machine in the Urban Design Loop

From Design Computation
Revision as of 11:42, 14 June 2022 by Abel Maciel (talk | contribs)

Jump to: navigation, search
DCIO2021-Logo.png
DC I/O 2021 Paper by STEPHEN LAW.
DCIO2021 S-Law.png

Abstract

We are living in an age where data is ubiquitous and data science methods are pervasive. The question is how do we adapt these methods into the urban design process. The aim of this opinion piece is to serve as a platform for discussion. Urban design is the design of towns and cities, streets and spaces and its process is interdisciplinary and iterative; first in analysing its context, second in developing designs and third in evaluating the design collectively among stakeholders then repeating this process. In this talk, we will illustrates examples on adapting data science methods into the urban design process; i. discover knowledge from urban imagery and 3D point cloud data to better understand the context, ii. the use of generative methods in design development and the use of socio-economic performance evaluation models on assisting design. We call this process machine-in-the-urban-design-loop. We will conclude by describing some potential outlooks and challenges for this integration; the lack of labelled detail design data (sequential data and design decision data) and the lack of interpretability, explanability and controllability in black-box machine learning models.

Presentation

Left Video Recording.

Conference Paper

Left Conference Paper.

Keywords

urban design, data science, machine learning, geographic data science, design computation.

Reference

DOI: https://doi.org/10.47330/DCIO.2021.DDJD8821

Bibliography

PENDING