Ghim Ping Raymond Ong

*Scaling Pedestrian Crossing Analysis to 100 US Cities* To inform pedestrian-infrastructure investments, we need data accurate down to the block level, and scaled up to entire cites. In order to do this, we leveraged an advanced computer-vision model (Meta's Segment Anything Model, or SAM) to measure crossing distance throughout the 100 largest cities in the United States. Our analysis: - Fetched roughly 3 million satellite images of urban intersections; - Trained SAM to differentiate between drivable and non-drivable surfaces within those images; - Overlaid crossing edges extracted from OpenStreetMap on those segmented images; and - Deployed a grow-cut algorithm to cut each crossing edge at the intersecting sidewalk. This resulted in automated crossing distances throughout each of the 100 largest cities, which expands the analytical scope to inter-city, regional, and even national. Indeed, we found universal patterns in crossing distance nationwide, as well as differences associated with region and year of incorporation, illustrating the increasing trend toward car-centric planning in a novel way. The full pre-print is up now at the link below, and we have posted all analysis steps and related code on Github. https://lnkd.in/exsd5EeK The team behind this study was a genuine joy to work with: recently-graduated MS in Data Science students Arunav Gupta and Jiali Qian, anchored by Professor Debra F. Laefer. Their creativity, perseverance, and enthusiasm for building safer streets turned a capstone project into a exciting advancement of AI in the realm of transportation planning and urban design. Thanks to the NYU Center for Urban Science + Progress and NYU Center for Data Science. #computervision #urbanai #urbanscience #preprint #dataviz #cities

712

39 Comments

Our latest publication in AFM led by ex-Masters student Arijit Chakraborty. 🔗 Full article: https://lnkd.in/gKnv6zWg → Implements a joint data assimilation framework within the Indian Land Data Assimilation System (ILDAS) to assimilate Soil Moisture Active Passive (SMAP) and GLASS leaf area index (LAI) data, to explore the influence on evapotranspiration (ET) and gross primary production (GPP). → DA significantly reduces the variability in soil moisture, and LAI estimates across most river basins → Winter seasons show higher variability (standard deviation 0.041 m3/m3) due to localized irrigation effects captured by SMAP data. → DA significantly improves GPP estimates, especially during the monsoon and post-monsoon seasons → DA improves ET estimates. 🏠 HydroSense Lab: https://lnkd.in/dhpmaVr Acknowledgements: Ministry of Earth Sciences | India Meteorological Department

282

4 Comments

What does it take to measure gender equity and embed it into transportation decision-making and research practices? Our new article in Journal of Transportation Engineering, Part A: Systems (ASCE Publishing) explores a global framework for addressing this critical question: https://lnkd.in/gqUaaFTT. Congratulations to Sushmita Bhandari for leading this immense review and framework development effort, and to Melrose Meiyu Pan for her valuable collaboration throughout this project! Understanding how to measure gender equity, and therefore design transportation systems that reflect diverse needs, is critical for building equitable and inclusive communities.

84

🎉 Great news! Our article “Digital Twin-Assisted Urban Resilience: A Data-driven Framework for Sustainable Regeneration in Paranoá, Brasília” by Tao Dong and Massimo Tadi has been accepted for publication in Urban Science. This work explores how digital twin technology can support sustainable urban regeneration and enhance resilience — combining data-driven insights with design strategies to tackle real-world challenges. Stay tuned for the official publication link! 🌍✨ #UrbanResilience #DigitalTwin #SustainableCities #UrbanScience #Sustainability #IMM

97

1 Comment