Facilitating the Spread of Knowledge and Innovation in Professional Software Development

Write for InfoQ


Choose your language

InfoQ Homepage News Sidewalk Labs Launches Machine Learning Tool for Urban Design

Sidewalk Labs Launches Machine Learning Tool for Urban Design

This item in japanese


Sidewalk Labs recently released Delve, a generative design tool powered by machine learning (ML), which helps developers, architects, and planners design urban neighborhoods. The ML algorithms can generate design concepts from minimal user input about the space and the goals of the project, while also measuring the impact of each design choice.

Delve works by creating and ranking urban desigs based off of criteria such as budget, location, and size. The tool can generate high-fidelity designs by allocating individual housing units across a project site and evaluating the daylight for each unit, and optimizing for open space access, walkability and views. Every design comes with a predicted cost per building, street section and open space. Sidewalk’s sustainability team built models for estimating and optimizing electricity use, water, waste, and rooftop solar intake.


A pre-launch version of Delve was deployed by Quintain, a U.K- and Ireland-based urban development team, for a recently completed mixed-use site at Wembley Park in North West London. 40,000 design variants were generated, evaluated, and optimized by Delve, and in the final set of variants Delve was able to exceed benchmark performance for key metrics on unit yield, daylight access, sun hours on the ground, and daylight impacts on neighbors. The case study explained, "Quintain had ambitious goals for the parcel and needed a design that could satisfy multiple spatial and performance requirements. Previous designs from 2016 and 2019 either did not provide sufficient unit yield to achieve the project’s financial objectives or reduced daylight access".

New York-based architect Sebastian Errazuriz predicted a year before the launch of Delve, that ninety percent of architects would lose their jobs as artificial intelligence takes over the design process. He described a theoretical job-killing app that could allow clients to describe what type of building they want, the budget, location, size, and other preferences and quickly get results. Further to Errazuriz’s prediction, there are other companies on the rise that are working on similar applications to Delve, like Spacemaker. Sidewalk Labs' director of product management Okalo Ikhena has a more optimistic view on these types of applications, stating that "Delve is providing developers and designers with superpowers!"

Rate this Article