About Us

The Network Data Science Laboratory at the School of Information Sciences at the University of Pittsburgh is lead by Konstantinos Pelechrinis. Research projects are targeted towards empirical and theoretical studies of networks and their appications in social, urban, technological, economic and biological networks.

Many of the interactions within the systems of interest can be captured through one or higher-mode networks. Our approach is data-driven and involves the development of novel methods, models and algorithms mainly focused on networks. Our research is driven by a variety of data, including civic (e.g., parking meter data, public transportation data, construction data etc.), social media (e.g., social networks, check-ins etc.) and computer network datasets (e.g., AS Internet topology, network flows etc.).

Latest Group News

Our paper on the economic impact of street fairs will be presented in ECML/PKDD 2016. Konstantinos Pelechrinis will be delivering a tutorial on urban informatics at SBP-BRiMS. Konstantinos Pelechrinis will be delivering a tutorial on the Web of Cities with Daniele Quercia, Anastasios Noulas and Bruno Goncalves at AAAI ICWSM 2016.
Our paper on spectral methods for analyzing mobility networks will be presented at AAAI ICWSM 2016.
New paper at PLOS ONE by our group, studying the fractal dimensionality of complex networks.
The group's latest publication at PLOS ONE develops a metric for quantifying multidimensional assortativity.
Our group will be presenting two papers in the Workshop on Information Networks to be held at NYU.
Dr. Pelechrinis received the Army Research Office Young Investigator award for "Models and Metrics for Composite Socio-Spatial Networks".
Our work on social media promotions will be presented at AAAI ICWSM. Media coverage: Pitt News, Pittsburgh Post Gazette, Pittsburgh's NPR News Station, Radio PA, Bloomberg Business, GeoMarketing, World News
Konstantinos Pelechrinis will be delivering a tutorial on Urban Informatics and the Web with Daniele Quercia at ACM WWW 2015.



Konstantinos Pelechrinis

Affiliated Faculty

Yuru Lin

Prashant Krishnamurthy

PhD Students

Ke Zhang

Anh Le

Dong Wei


Theodoros Lappas (Stevens Institute of Technology)

Evangelos Papalexakis (Carnegie Mellon University)

Evimaria Terzi (Boston University)

Michalis Faloutsos (University of California, Riverside)

Christos Faloutsos (Carnegie Mellon University)

Marios Kokkodis (Boston College)

Research Projects in a Nutshell

Methods and Metrics for Composite Complex Networks

The vast majority of existing work in network analysis deals with unimodal networks and cannot flexibly capture the heterogeneity of nodes and/or edges of a composite network. An approach that is typically followed is to analyze single-mode projections of the composite network on one of the node types, which is associated with a significant amount of information loss. Flexible ways to discover and mine hidden patterns is still an open and challenging problem for composite networks. In addition, traditional graph metrics that quantify network properties cannot be applied on composite networks. Hence, in this project we aim to fill this gap in the literature by developing (i) a composite network modeling framework using tensors and (ii) network metrics for mining composite networks.

Analyzing and Modeling Location-based Social Networks

The advancements of mobile computing technologies has lead to a new class of social media, namely, location-based social networks. The main interaction in these platforms is sharing/broadcasting location, thus, tying the digital and the physical world. In this project, our goal is to analyze and model the underlying network structures (e.g., friendship graph, user-location affiliation graphs, location co-visitation graphts etc.) to study phenomena such as homophily, peer influence, social selection and the structure of the local economy ecosystem. Of particular interest is understanind and modeling the effects of external shocks (e.g., local government decisions) on the underlying social and urban network system.

Selected Publications






The Network Data Science Laboratory offers every year (Spring semester) an advanced class on Network Science and Analysis

Prior offerings of the class can be found here:

Doctoral seminars are also offered periodically:

  • Doctoral Seminar on Social Networks and Graph Analysis (Lin - Spring 2015)
  • Doctoral Seminar on Data Science (Lin - Fall 2014)
  • Doctoral Seminar on Location Based Social Networks (Pelechrinis - Spring 2012)

The Network Data Science Lab participates in various educational outreach activities: