https://www.zhuangyuanfan.com/about daily 0.75 2025-07-11 https://images.squarespace-cdn.com/content/v1/5afb96837106998707f9b91a/1550350583616-OUNFVSTCYHWL3ANSLZSE/dibao.jpg About https://images.squarespace-cdn.com/content/v1/5afb96837106998707f9b91a/1550350583616-OUNFVSTCYHWL3ANSLZSE/dibao.jpg About https://www.zhuangyuanfan.com/contact daily 0.75 2026-09-07 https://images.squarespace-cdn.com/content/v1/5afb96837106998707f9b91a/1550345483707-T67VN9L6BTWG1CMXVO9Y/image-asset.jpeg Contact https://images.squarespace-cdn.com/content/v1/5afb96837106998707f9b91a/1550345483707-T67VN9L6BTWG1CMXVO9Y/image-asset.jpeg Contact https://www.zhuangyuanfan.com/work daily 1.0 2026-09-25 https://www.zhuangyuanfan.com/work/covid19-and-cities-lg788 monthly 0.5 2026-09-07 https://images.squarespace-cdn.com/content/v1/5afb96837106998707f9b91a/1592931577499-2CE1GDTAJGU46VU69ZEJ/mapforweb1.png Projects - Covid-19 and Cities https://images.squarespace-cdn.com/content/v1/5afb96837106998707f9b91a/1592931599966-3EW7ODPZSOIA0BG21RHQ/image4.png Projects - Covid-19 and Cities https://www.zhuangyuanfan.com/work/perception-bias-52wbc monthly 0.5 2026-09-07 https://images.squarespace-cdn.com/content/v1/5afb96837106998707f9b91a/1608288665242-E5M7RA3J28VBQJ4DGWXZ/final_GSV-11.jpg Projects - Perception Bias https://images.squarespace-cdn.com/content/v1/5afb96837106998707f9b91a/1608289120382-BLSNVNEMDLC43VKLY7MX/site-09.jpg Projects - Perception Bias https://images.squarespace-cdn.com/content/v1/5afb96837106998707f9b91a/1608289148544-PA57DSMCME02TI5BIA7N/perceived+safety-11.jpg Projects - Perception Bias https://images.squarespace-cdn.com/content/v1/5afb96837106998707f9b91a/1608289433675-9YQFGBVV19N6LKCQ0JLC/final_researchDesign.jpeg Projects - Perception Bias https://www.zhuangyuanfan.com/work/desirable-streets-rc796 monthly 0.5 2026-09-07 https://images.squarespace-cdn.com/content/v1/5afb96837106998707f9b91a/1623725891473-XL95VIKC7C0IDN4IACF9/TitleCard.jpg Projects - Desirable Streets Paper published at Computers, Environment and Urban Systems Interactive Map Here The experience of walking through a city is influenced by amenities and the visual qualities of its built environment. This paper uses thousands of pedestrian trajectories obtained from GPS signals to construct a desirability index for streets in Boston. We create the index by comparing the actual paths taken by pedestrians with the shortest path between any origin-destination pairs. The index captures pedestrians’ willingness to deviate from their shortest path and provides a measure of the scenic and experience value provided by different parts of the city. We then use computer vision techniques combined with georeferenced data to measure the built environment of streets. We show that desirable streets have better access to public amenities such as parks, sidewalks, and urban furniture. They are also sinuous, visually enclosed, have less complex facades, and have more diverse business establishments. These results further our understanding of the value that the built environment brings to pedestrians, enhancing our capacity to design more lively and functional environments. Arianna Salazar Miranda, Zhuangyuan Fan, Fabio Duarte, Carlo Ratti, “Desirable Street: Using Deviations in Pedestrian Trajectories to Measure the Value of the Built Environment”, Computers, Environment and Urban System ,https://doi.org/10.1016/j.compenvurbsys.2020.101563. https://images.squarespace-cdn.com/content/v1/5afb96837106998707f9b91a/1623725893389-4LSNK5WJMQDZOIUDVFVT/LechmerePath.jpg Projects - Desirable Streets https://www.zhuangyuanfan.com/work/o2o-wfy5w monthly 0.5 2026-09-07 https://images.squarespace-cdn.com/content/v1/5afb96837106998707f9b91a/1623726373809-EC2BKSPCTATF3Z9FYOL9/Screen+Shot+2021-06-15+at+10.54.45+AM.png Projects - Online Offline Connection https://images.squarespace-cdn.com/content/v1/5afb96837106998707f9b91a/1623726389349-0P7Y26713L3VU7RZG0HI/total_num_review_raw_visit_counts_all.jpg Projects - Online Offline Connection https://www.zhuangyuanfan.com/work/rhythm-of-transit-stations-e6mf8 monthly 0.5 2026-09-07 https://images.squarespace-cdn.com/content/v1/5afb96837106998707f9b91a/1640400870735-EKF0HYQG5ZNURCUTHZ95/graphics_rev-02.png Projects - Rhythm of Transit Stations https://images.squarespace-cdn.com/content/v1/5afb96837106998707f9b91a/1640400499986-NXT746PWAG4MVFBP4ZQP/graphics_rev-01.png Projects - Rhythm of Transit Stations https://images.squarespace-cdn.com/content/v1/5afb96837106998707f9b91a/1640400631026-WZUJGAE4BKDDVP9H5LJM/tod_research2.png Projects - Rhythm of Transit Stations https://images.squarespace-cdn.com/content/v1/5afb96837106998707f9b91a/1640400891076-GPID3K02H1KSBV6RN8V7/graphics_rev-05.png Projects - Rhythm of Transit Stations https://images.squarespace-cdn.com/content/v1/5afb96837106998707f9b91a/1640400924172-MUSHMPNRZTSA7YXI7KK0/graphics_rev-10.png Projects - Rhythm of Transit Stations https://www.zhuangyuanfan.com/work/jaywalking-grsa5 monthly 0.5 2026-09-07 https://images.squarespace-cdn.com/content/v1/5afb96837106998707f9b91a/1649319212464-79I8QB65E9O8HQPIAZVD/jaywalking+heatmap.gif Projects - Jaywalking in Cities - Dynamic Jaywalking Detected in Hong Kong (Copy) Dynamic Jaywalking Detected in Hong Kong https://images.squarespace-cdn.com/content/v1/5afb96837106998707f9b91a/1649319810746-JODMC76O7UMWE5T0977H/datasource_method_Fig+8.png Projects - Jaywalking in Cities https://www.zhuangyuanfan.com/work/food-battle-swra3 monthly 0.5 2026-09-07 https://images.squarespace-cdn.com/content/v1/5afb96837106998707f9b91a/1661503987862-P5OHSC0Q8H62H9B0PIK4/foodbattle.png Projects - Food Battle https://www.zhuangyuanfan.com/work/great-streets-3txat monthly 0.5 2026-09-07 https://images.squarespace-cdn.com/content/v1/5afb96837106998707f9b91a/1712644912270-REMYL8U56RC8V06KUXPW/website_snapshot.jpg Projects - Great Streets - The Great Streets (Copy) The Great Streets https://images.squarespace-cdn.com/content/v1/5afb96837106998707f9b91a/1712644698738-BCG7NVOI4JI3D3EZ4V0T/Fig_Fig2-ESMExamples.png Projects - Great Streets https://images.squarespace-cdn.com/content/v1/5afb96837106998707f9b91a/1712644709914-RNWL7C230ZWBN3DELI7H/Fig_Fig1-ESM.png Projects - Great Streets https://www.zhuangyuanfan.com/work/urban-visual-intelligence-mdc9n monthly 0.5 2026-09-25 https://images.squarespace-cdn.com/content/v1/5afb96837106998707f9b91a/1692175509452-JXLWON10URFSLM2ZPHM6/vision-natural.jpeg Projects - Urban Visual Intelligence Image Credit: Sebastian Meier, with support from Till Nagel and the MIT Senseable City Lab. https://images.squarespace-cdn.com/content/v1/5afb96837106998707f9b91a/1692175275120-AOS48OV32DKEXEUOZJP2/Fig_1-ResearchStructure.png Projects - Urban Visual Intelligence https://images.squarespace-cdn.com/content/v1/5afb96837106998707f9b91a/1692175577952-1VBY69E2RUZPGD922KIP/Fig_2-Q1result_2-Q1result.png Projects - Urban Visual Intelligence https://www.zhuangyuanfan.com/work/urban-visual-clusters-4gbep monthly 0.5 2026-09-07 https://images.squarespace-cdn.com/content/v1/5afb96837106998707f9b91a/1747272369530-83N093ISY72D31ECUXAQ/GIF+EDITS+Urban+Visual+Clusters+and+Road+Transport+Fatalities.gif Projects - Urban Visual Clusters https://images.squarespace-cdn.com/content/v1/5afb96837106998707f9b91a/1746433592390-MRQMYKBUT1T4HJ8Q3651/Fig1-02.png Projects - Urban Visual Clusters https://images.squarespace-cdn.com/content/v1/5afb96837106998707f9b91a/1746433382981-4XROEJV5C616KKUUKVMI/fig3.png Projects - Urban Visual Clusters https://images.squarespace-cdn.com/content/v1/5afb96837106998707f9b91a/1746433612082-GYDLC07YI5NLO43OED7X/Fig2-01.png Projects - Urban Visual Clusters https://images.squarespace-cdn.com/content/v1/5afb96837106998707f9b91a/1747272347531-B2J169IHAUSE6KRUUAPC/GIF+EDITS+Urban+Visual+Clusters+and+Road+Transport+Fatalities-p1.gif Projects - Urban Visual Clusters https://www.zhuangyuanfan.com/work/social-mixing-in-five-global-cities monthly 0.5 2026-09-25 https://images.squarespace-cdn.com/content/v1/5afb96837106998707f9b91a/d8a06b93-ff1a-42a3-9def-3c203198e573/Fig_fig2.png Projects - Hidden patterns of urban mixing across five global cities Explain daytime social mixing. a. DM is lower for low-income groups than for mid-income groups in Boston (n= 5,894), Chicago (n = 15,738), and São Paulo (n = 40,896), while gaps are less pronounced in Hong Kong (n = 51,619) and London (n = 30,277). Only people with age>17 are included. b. DM distribution by age and city. c. Home accessibility to train stations by income group in each city. d. Home accessibility to bus stops by household car ownership. e. Average DM across all cities by age groups and work status. DM declines gradually with age (dashed line), but rises again after retirement (65–74). Working individuals report higher DM than non-working peers, especially at younger ages. f. DM distribution by gender and care-taking responsibilities at early career age. Only person between age 21 to 45 are included (Boston n = 2,064, Chicago n = 8,365, Hong Kong, n = 25,594, London, n = 15,394, São Paulo, n = 20,227). In a-d, f, the solid dot marker shows the mean value of the sample by each group. The error bar area represents the 95% confidence interval. https://images.squarespace-cdn.com/content/v1/5afb96837106998707f9b91a/e337065c-ae08-43d6-899c-afa64f1f76c5/Fig_Fig2-place-exposure.png Projects - Hidden patterns of urban mixing across five global cities Conceptual diagram of an encoder-decoder architecture that predicts individual place exposure. a. Construct a POI place spatial network using distance and travel-time based method. Then use a GNN model, we encode all POI places in each city into an embedding vector. For each city, we have M POIs and the embedding dimension is 32. b. Create the feature embedding. For each person’s home location, we construct a 15-minute walking isochrone and gets all available POI places within the isochrone. The home-space embedding (h_{h}) is the mean embedding of all POI places associated with their home location 15-minute walking isochrone. Similarly, for each person, we got their maximum travel activity space, a polygon that contains all locations that the person ever visited during the day. The activity-space embedding (h_{a}) is the mean embedding of all POI places within the person’s activity space. Sociodemographic embedding includes all personal attributes such as age, gender, and work status. c. The autoencoder model takes the three embedding in a permutation fashion and compress them into a latent code z, and then decode to an output layer, which is the N-Dimensional POI place exposure vector (N is the total POI types). https://images.squarespace-cdn.com/content/v1/5afb96837106998707f9b91a/280cd4c2-ec6c-4898-93a7-2433799e0ab3/fig5-01.png Projects - Hidden patterns of urban mixing across five global cities Predicting individual-level place exposure. a. The mean R2 of models that include h_{h},h_{h}\left|\left|h_{d},h_{a},h_{a}\right|\right|h_{d}. || indicates that two embeddings are further concatenated to create a latent code for the decoder to reconstruct the place exposure. b. We use the Low-income group as the training data and train the autoencoder model and then apply the model to predict the place exposure of all other income groups. ML stands for Medium Low; MH stands for Medium High; H stands for High. The dash line indicates the model performance using random split stratified on income (75% data used for training and 25% used for testing). https://images.squarespace-cdn.com/content/v1/5afb96837106998707f9b91a/9c75ab79-9ecc-48e2-b52e-272a04fe0b6c/Fig_Fig1-selected.png