William H. Aeberhard

Senior Data Scientist

Swiss Data Science Center

About me

I obtained my PhD in Statistics in 2015 in a cotutelle between the University of Geneva and the University of Sydney. I then worked as a post-doctoral fellow at Dalhousie University as part of a Canadian Statistical Sciences Institute collaborative research team. I was an Assistant Professor of Statistics and Coordinator of the MS in Data Science in the Department of Mathematical Sciences at Stevens Institute of Technology before joining the Swiss Data Science Center, affiliated with ETH Zurich, in 2020 as Senior Data Scientist.

My research interests lie in robust statistics, semi-parametric methods, computational statistics, and spatio-temporal modeling with ecological applications. My current methodological research includes robust smoothing parameter selection for non-parametric additive models and implementations of outlier detection for general state space models. Ongoing cross-disciplinary collaborations include growth estimation methods from mark-recapture data and large-scale spatio-temporal modeling of the impact of volcanic eruptions on vegetation density from satellite imagery.


  • Robust Statistics
  • Non- and Semi-Parametric Methods
  • Spatio-Temporal Modeling with Ecological Applications


  • PhD in Statistics, 2015

    University of Geneva and University of Sydney

  • MSc in Statistics, 2010

    University of Geneva

  • BSc in Psychology, 2008

    University of Geneva

Published Papers

(2018). Review of State-Space Models for Fisheries Science. Annual Review of Statistics and Its Application 5, 215–235.


Book Chapters

(2015). Le Modèle Linéaire Généralisé (GLM) Robuste. Méthodes Robustes en Statistique. Ed. by J.-J. Droesbeke, G. Saporta, and C. Thomas-Agnan. Paris, France: Technip, 109–128. ISBN: 9782710811497.



R code for robust estimation for state space models, supplementary material to Aeberhard, Cantoni, Field, Kuensch, Mills Flemming and Xu (2019)


R package for fitting of semi-parametric generalized linear models based on exponentially tilted mixtures


R code for simulating and fitting the alternative state space model of Yin, Aeberhard, Smith and Mills Flemming (2018)


Fully-documented R package for truncated negative binomial and truncated Poisson family objects


Fully-documented R package for robust estimation and accurate inference for the negative binomial model, implementing the methods presented in Aeberhard, Cantoni and Heritier (2017)


R code for fitting a robust negative binomial generalized linear model, supplementary material to Aeberhard, Cantoni and Heritier (2014)


Lecturer at Stevens Institute of Technology

  • MA 540 Introduction to Probability Theory: probability theory for MSc in Data Science (Spring ‘20)
  • MA 611 Probability: measure-theoretic probability theory for MSc and PhD in Mathematics (Fall ‘18, Fall ‘19)
  • MA 701 Statistical Inference: fundamental statistics for PhD in Business Administration and other MSc/PhD programs (Fall ‘19)
  • MA 331 Intermediate Statistics: basics of statistical inference for 2nd year BSc in many science and engineering curricula (Spring ‘19)

Lecturer at Dalhousie University

  • 1060 Introductory Statistics for Science and Health Science: basics of explorartory data analysis and statistical inference for 1st year BSc in health sciences (Summer ‘16, Winter ‘17)