Where did the idea for FaultLines originate?

It has been more than a year since two powerful earthquakes struck the southeast of Turkey, taking more than 60,000 lives overnight and causing $34.2 billion in damage. Despite being some of the largest earthquakes recorded in the region’s history, multiple reports attribute the scale of damage to widespread construction malpractices. These malpractices, orchestrated by corrupt officials in the construction industry, result in a common set of structural defects that are visible even to the bare eye. More pressingly, seismologists predict there’s a 70% chance that an equally powerful earthquake will hit Istanbul in the next five years, a city where 20 million people live in 1.2 million buildings, more than half of which are expected to sustain significant damage — if not collapse completely. At the current rate of municipal structural testing — 120 buildings per day — it may take up to 21 years for a resident to see if their building is safe. Hence I came up with an idea to look for these visible defects using computer vision and image segmentation algorithms (AI), across millions of Google StreetView Images. 

This idea got me a research position at Columbia University’s Center for Spatial Research, under the supervision of Laura Kurgan who’s been immensely supportive of me and my studies here at GSAPP - and I’m very happy that the University’s also supporting me now with this grant and an incredible opportunity to further develop this idea into a tool that could be very useful! 

What is FaultLines?
Faultlines is an AI-powered Earthquake Risk Assessment Platform.

FaultLines, the culmination of a year-long research project funded by Columbia GSAPP and the Center for Spatial Research, aims to use novel machine learning models to detect structural defects based solely on Google StreetView images of buildings, providing critical safety insights in a matter of seconds.

How does it work?
Our method starts with gathering various geospatial, municipal, and seismic data to create a well-rounded base map of any given city. We then use semantic image segmentation models to detect structural defects across millions of Google StreetView images and generate a risk score for each building.

Cumulatively, this enables us to create an urban-scale, interactive digital map that anyone with an internet connection can access. Users simply type in their address to look up their building’s risk score, get a detailed report behind our insights, and see what actions and safety measures they can take next.

Our platform will also offer large-scale data access to public organizations, such as governments and municipal agencies, as well as private enterprises, especially those in real estate, construction, insurance, and financial services. Through data licensing and API access, municipalities, who are currently challenged by blindly testing millions of structures, can optimize their resources by testing according to anticipated risk — allowing them to focus on the most vulnerable buildings first, and shortening waitlists significantly. Real estate businesses, such as Zillow, could directly integrate FaultLines into their platform by providing risk reports for their listings, allowing potential renters, buyers, and investors to make more informed decisions.

Ultimately, our project aims to further develop a novel AI method to get useful insights about the built environment, generating large-scale datasets across 110+ countries and regions where street imagery data is available. By harnessing this data into an interactive digital platform, we can improve safety assessments and decision-making in urban environments worldwide, and through large-scale data licensing and API access we can offer great value to businesses across a variety of sectors.

How can supporters follow the project?
The Brown Institute’s blog will document the trajectory of winning projects over the next few months.