The Livehoods Project: Utilizing Social Media to Understand the Dynamics of a City

Date

July 30, 2012

Speaker

Justin Cranshaw

Affiliation

Carnegie Mellon University

Overview

The forces that shape the dynamics of a city are multifarious and complex. Cultural perceptions, economic factors, municipal borders, demography, geography, and resources—all shape and constrain the texture and character of local urban life. Studying the intricate and rapidly-evolving sociocultural structure of the city has traditionally been a challenging endeavor for researchers, often requiring hundreds of hours of observation and interviews. Although such methods offer a way to gather deep insights about certain aspects of city life, they simply do not scale, and so can ever only uncover a partial image of the inner workings of the city.

The Livehoods project (http://livehoods.org) presents a new and promising methodology for studying the dynamics, structure, and character of a city on a large scale. Our approach is fundamentally data-driven. Using geospatial social media data such as Tweets and Check-ins, we use clustering algorithms to discover the hidden structures of the city. The discovered clusters, which we call Livehoods, reveal a snap-shot of the dynamic areas that comprise the urban environment. Like neighborhoods, Livehoods are a representation of the organizational structure of the city. However, Livehoods are different from neighborhoods. Livehoods allow us to investigate and explore how people actually use the city, simultaneously shedding light onto the factors that come together to shape the urban landscape and the social texture of city life. In this talk we will present and overview of our approach and we discuss results of our field study exploring and validating the Livehoods of Pittsburgh.

This work was done in collaboration with Norman Sadeh, Jason Hong, and Raz Schwartz.

Speakers

Justin Cranshaw

Justin Cranshaw is a PhD student in the School of Computer Science at Carnegie Mellon University. His research combines methods from machine learning, human-computer interaction, and urban studies, to investigate how emerging large-scale sources of rich and expressive social media data can be used as a lens to help us better understand the structure, social dynamics, and character of cities.

People