Related Experiment Video
Updated: Jan 25, 2026

The Social Dimension of Stress: Experimental Manipulations of Social Support and Social Identity in the Trier Social Stress Test
Published on: November 19, 2015
Discovering place-informative scenes and objects using social media photos
Fan Zhang1,2, Bolei Zhou3, Carlo Ratti2
1Institute of Remote Sensing and Geographical Information Systems, School of Earth and Space Sciences, Peking University, Beijing 100871, People's Republic of China.
This study introduces a data-driven framework using deep learning to analyze visual place characteristics from millions of geotagged photos. It identifies key visual elements like historical architecture and unique urban scenes that define city appearance.
Area of Science:
- Computer Science
- Urban Studies
- Architectural Design
Background:
- Quantifying place's visual characteristics is crucial for urban design and tourism.
- Previous research faced limitations due to insufficient data and methods for visual analysis.
Purpose of the Study:
- To develop a data-driven framework for exploring place-informative scenes and objects.
- To automatically learn and measure visual place appearance from large photo datasets.
- To compare visual similarity and distinctiveness across global cities.
Main Methods:
- Utilized a deep convolutional neural network (CNN) for visual knowledge extraction.
- Analyzed millions of geotagged photos from social media for 18 cities worldwide.
- Employed a data-driven framework to identify place-informative visual cues.
Main Results:
- Identified historical architecture, religious sites, unique urban scenes, and natural landscapes as key place-informative elements.
- Found specific vehicles like taxis, police cars, and ambulances to be highly city-informative.
- Revealed visual cues that distinguish cities beyond traditional landmarks.
Conclusions:
- Large-scale geotagged data, analyzed via deep learning, offers significant insights into place formalization and urban design.
- The framework provides a novel method for understanding and quantifying urban visual identity.
- Findings are applicable to architectural design, urban planning, and tourism.
Related Concept Videos
Social Proof
Social Scripts
Social Traps
Social Facilitation
Social Exchange Theory
Social Loafing

