Current Position

I am presently a Senior Lecturer (in the Commonwealth usage) in the Department of Statistics at the School of Mathematics and Statistics of University of New South Wales, in Sydney.

Contents

This site has information on my academic writing; presentations I have made; scientific software on which I have worked; and my full academic CV.

Current Research Topics

Generative models for networks

I work in development and efficient implementation of generative models for social networks and social network processes, with applications in epidemiology and the social sciences.

Dynamic networks

Models for networks that evolve over time have manifold application in areas as diverse as epidemiology, social sciences, and marketing. A major component of my research is the development of realistic yet parsimonious models for the evolution of social networks, and fitting them based on available data. These models have been applied to the modeling and simulation of disease spread.

Exponential-family random graph models

Exponential-family random graph models provide a rigorous and flexible framework for modeling network structure and evolution, but have the disadvantage of being computationally intensive to fit. My research interest relating to these models is extending them to valued and rank data, diagnostics, and efficient implementation that takes advantage of their conditional dependence structure.

Partially and egocentrically observed networks

Many real-world networks, whether social or epidemic, are not practical to observe directly. Instead, a network may be observed indirectly, from the point of view of a sample of individuals in the network. Or, only some of the relations may be observed, and noisily at that. My work develops methodology and software to permit network structure to be recovered from such indirectly observed relational data.

Latent space models

A powerful approach to modeling complex network structure is through latent variable models, in which an unobserved (latent) social space in which actors interact and/or unobserved attributes of actors are postulated that account for observed network structure. My interest is in extending them to dynamic network data, applying them to voting data, and efficient Bayesian inference.

“Big Data” networks

Block chain transactions

Transactions on the block chain—such as BitCoin—are, in principle, anonymous, but also transparent. What can be learned?

Telecommunications networks

Network data collected by telecommunication servies such as mobile carriers are characterized by their massive scale — millions of individuals — and their great degree of granularity, recording every call and SMS message.