This is an old revision of the document!

CityPulse: Understanding Urban Spatio-Temporal Pattern from Mobile Activity


  • Understand city dynamics (applications: smart cities, urban computing)
  • Automated characterization of city regions through mobile phone usage
  • Predict future activity based on past spatio-temporal activity (application: city, network planning)


Our goal is to :

  • Use data from mobile activity (e.g. calls, SMS, Internet)
  • Identify and interpret spatio-temporal patterns (e.g. clusters)
  • Do it efficiently (design computationally efficient algorithms)

Video of heatmap activity for 24h in Milan - Nov.1st 2013
Dataset “Telecommunications – SMS, Call, Internet – MI”:


  • Use temporal activity as a “signature” of a grid square:
    • Activity = Calls, SMS, and Internet data
  • Use this signature to characterize the area


Used data

  1. Telecommunications – SMS, Call, Internet – MI :
    • We aggregate Call/SMS activity, over all country codes, for each timestamp
    • We aggregate Internet activity over all country codes
    • We normalize all time series, by removing the mean, and dividing with the standard deviation.
  2. Telecommunications MI-to-MI
  3. Open data collected from the city of Milano
    • Tourism: Accommodations hotels (2013)
    • Museums and exhibition spaces in Milan
    • Sports, cinemas, etc.


Blerim Cici, Ph.D. candidate, Information and Computer Science Department
Xuhong Zhang , Ph.D. student, Electrical Engineering and Computer Science Department
Minas Gjoka, Postdoc, Calit2
Carter T. Butts, Professor, Sociology, Statistics, and EECS Departments
Athina Markopoulou, Associate Professor, Electrical Engineering and Computer Science Department

Research Group Websites:
Networking Group:
Network, Computations, and Social Dynamics Group:

Contact Person: Blerim Cici (bcici AT uci DOT edu)