Project

CityMatrix

Copyright

Yan Zhang 'Ryan', City Science, MIT Media Lab

Yan Zhang 'Ryan', City Science, MIT Media Lab


Thesis Citation

Zhang, Yan. “CityMatrix – An Urban Decision Support System Augmented by Artificial Intelligence.” Massachusetts Institute of Technology, 2017.


An Urban Decision-Support System Augmented by Artificial Intelligence

The decision-making process in urban design and urban planning is outdated. Currently, urban decision-making is mostly a top-down process, with community participation only in its late stages. Furthermore, many design decisions are subjective, rather than based on quantifiable performance and data. Current tools for urban planning do not allow both expert and non-expert stakeholders to explore a range of complex scenarios rapidly with real-time feedback. 

CityMatrix was an effort towards evidence-based, democratic decision-making. Its contributions lie in the application of Machine Learning as a versatile, quick, accurate, and low-cost approach to enable real-time feedback of complex urban simulations and the implementation of the optimization searching algorithms to provide open-ended decision-making suggestions. The goals of CityMatrix were: 


  1. Designing an intuitive Tangible User Int… View full description


Thesis Citation

Zhang, Yan. “CityMatrix – An Urban Decision Support System Augmented by Artificial Intelligence.” Massachusetts Institute of Technology, 2017.


An Urban Decision-Support System Augmented by Artificial Intelligence

The decision-making process in urban design and urban planning is outdated. Currently, urban decision-making is mostly a top-down process, with community participation only in its late stages. Furthermore, many design decisions are subjective, rather than based on quantifiable performance and data. Current tools for urban planning do not allow both expert and non-expert stakeholders to explore a range of complex scenarios rapidly with real-time feedback. 

CityMatrix was an effort towards evidence-based, democratic decision-making. Its contributions lie in the application of Machine Learning as a versatile, quick, accurate, and low-cost approach to enable real-time feedback of complex urban simulations and the implementation of the optimization searching algorithms to provide open-ended decision-making suggestions. The goals of CityMatrix were: 


  1. Designing an intuitive Tangible User Interface (TUI) to improve the accessibility of the decision-making process for non-experts. 
  2. Creating real-time feedback on multi-objective urban performances to help users evaluate their decisions, thus to enable rapid, collaborative decision-making. 
  3. Constructing a suggestion-making system that frees stakeholders from excessive, quantitative considerations and allows them to focus on the qualitative aspects of the city, thus helping them define and achieve their goals more efficiently.

CityMatrix was augmented by Artificial Intelligence (AI) techniques including Machine Learning simulation predictions and optimization search algorithms. The hypothesis explored in this work was that the decision quality could be improved by the organic combination of both strengths of human intelligence and machine intelligence.

The system was pilot-tested and evaluated by comparing the problem-solving results of volunteers, with or without AI suggestions. Both quantitative and qualitative analytic results showed that CityMatrix is a promising tool that helps both professional and non-professional users understand the city better to make more collaborative and better-informed decisions. 

Copyright

All rights reserved by 'Ryan' Yan Zhang

Copyright

All rights reserved by 'Ryan' Yan Zhang

Copyright

All rights reserved by 'Ryan' Yan Zhang

Copyright

Yan Zhang 'Ryan', City Science, MIT Media Lab

Contact: Yan Zhang <ryanz@media.mit.edu>