Multimodal data
Video and mobile sensor data provide complementary perspectives on urban environments.
Urban Informatics · Accessibility · Citizen Science
Understanding urban accessibility.
SideSeeing explores urban accessibility through multimodal data, computer vision and machine learning. Discover research datasets from Brazil and the United States, and examine their images and predictions with DataViz.
Accessibility research
Computer vision and machine learning help investigate sidewalks and the conditions of everyday journeys.
Visual exploration
DataViz lets you examine frames, bounding boxes and segmentation masks within each collection.
People & institutions
Research behind SideSeeing
A collaborative project connecting computer vision, urban informatics and citizen science.
University of São Paulo (USP)
Institute of Mathematics and Statistics (IME)
São Paulo, Brazil
SideSeeing investigates how multimodal data can support the understanding of urban accessibility. Its research spans data collection, preprocessing, labeling, visualization, artificial intelligence and analysis for decision-making.
Rafael J. P. Damaceno
University of São Paulo (USP)
Fabio Miranda
University of Illinois at Chicago
Department of Computer ScienceMaryam Hosseini
Massachusetts Institute of Technology (MIT)
Department of Urban Studies and PlanningRoberto M. Cesar-Jr
University of São Paulo (USP)
Research credits and affiliations follow the official SideSeeing team page at USP .
Research datasets
Available collections
Each collection has its own viewer, with separate data and preferences.