Our latest study in London, UK, examines an innovative framework to quantify and map how pedestrians perceive urban greenery, utilizing advanced space syntax metrics and crowdsourced imagery for more informed urban planning and design.

Nature AI Lab
2025-02-15

We are glad to share our new paper:
Mingze Chen, Yuxuan Liu* , Fan Liu, Trishla Chadha, Keunhyun Park. Measuring pedestrian-level street greenery visibility through space syntax and crowdsourced imagery: A case study in London, UK. Urban Forestry & Urban Greening 2025 | https://doi.org/10.1016/j.ufug.2025.128725.[download]

🚀 A huge thanks to Said Turksever and Edoardo Neerhut from Meta for data support!

🔍 Why is this important?
Google Street View (GSV) has been widely used in urban greenery studies, but how can we assess street greenery from a true pedestrian perspective (rather than a vehicle-centric view)? Our study introduces a novel Pedestrian Green View Index (PGVI) and integrates:
1️⃣ Pedestrian crowdsourced imagery from Meta’s platform,
2️⃣ Rhino-GH-based tree canopy modeling and Visibility Graph Analysis (VGA) from space syntax,
3️⃣ Perception surveys with 183 volunteers covering 7 key urban experience dimensions.

🌳 What did we find?
We identified nine pedestrian greenery types in the City of London and compared traditional GVI, PGVI, VGA-GVI, and human perception data. Results show that PGVI aligns more closely with actual pedestrian experience, revealing the limits of space syntax alone in capturing nuanced human-nature interactions.

🎯 Key Takeaways for Urban Design & Planning:
✔️ Integrating multiple methods enhances our understanding of greenery perception.
✔️ Pedestrian-level analysis is crucial for designing inclusive, walkable, and greener cities.
✔️ Combining spatial analytics with human feedback leads to more effective urban planning strategies.