2018 Featured Speaker: Guilherme J. M. Rosa
University of Wisconsin-Madison
Titled: Harnessing Operational Farm Data to Enhance Advancement of Agricultural Sciences
Randomization is a powerful tool to investigate causal effects, and randomized experiments are extensively used in agricultural research. Quite often, however, randomization is not feasible. In addition, results of controlled agricultural experiments may not directly translate to real farm conditions. An alternative in this context is to harness the enormous amount of farm-recorded data routinely collected in agricultural and livestock operations. Inferring causal effects from such observational data is complex due to potential confounding effects and selection bias, so that a rigorous conceptual framework and careful analytic implementation of specific statistical methods and data mining techniques is required. Furthermore, the size and complexity of operational data poses many challenges in terms of data acquisition, data management, and analysis strategies. In this presentation we will discuss the potential utility of operational farm data to study causal effects in agriculture, with a focus on graphical modeling approaches. Key basic concepts of graphical models will be reviewed, and some examples using livestock data will be used to illustrate the methods.
About the Presenting Author
Guilherme Rosa obtained an M.S. in Animal Sciences from Sao Paulo State University (UNESP) – Brazil in 1994, and a Ph.D. in Statistics and Agricultural Experimentation from the University of Sao Paulo (USP) – Brazil in 1998. Guilherme Rosa started his professional career as a faculty member of the Department of Biostatistics at UNESP (1994-2001), then moved to the USA as a faculty member at Michigan State University (2002-2006), and is currently a Professor at the Department of Animal Sciences and the Department of Biostatistics & Medical Informatics at the University of Wisconsin-Madison (since 2006).
Guilherme Rosa teaches courses and develops research on quantitative genetics and statistical genomics, including design of experiments and data analysis tools. Some specific areas of interest include mixed effects models, Bayesian analysis, and Monte Carlo methods. More recently, Guilherme has been working also on the analysis of observational data in agriculture, using a variety of tools based either on the framework of potential outcomes, or graphical model tools, such as propensity score, instrumental variable, and Bayesian networks.
Guilherme has published 10 book chapters and over 150 refereed papers in scientific journals and has funded his program with outside grants valued at over $10 million. He has been awarded with the LeClerg Rotary Lecturer from the Biometrics Program at the University of Maryland (2011) and the Pond Research Award from the University of Wisconsin-Madison (2013). He has also received the Rockefeller Prentice Memorial Award in Animal Breeding and Genetics, by the American Society of Animal Science (2016), the Vilas Faculty Mid-Career Investigator Award from University of Wisconsin-Madison (2017), and the Excellence in International Activities Award, also from University of Wisconsin-Madison (2017).