Advancing in Statistical Modelling for evolutionary biologists and ecologists using R

Start: 
Mon, 2017/04/24 - Fri, 2017/04/28
Location: 
Loch Lomond, Scotland

This course will provide an introduction to working with real-life data typical of those encountered in the field of evolutionary biology and ecology. The course will be delivered by Dr. Luc Bussiere and Dr. Ane Timenes Laugen who are both practicing academics in the field of ecology.  This five day course will consist of series of modules (each lasting roughly half a day) covering model selection and simplification, generalised linear models, mixed effects models,  and non-linear models. Along the way you will gain in depth experience in data ‘wrangling’, data and model visualisation and plotting, as well as exploring and understanding model diagnostics. Classes will comprises of a mixture of lectures and practicals designed to either build required skills for future modules or to perform a family of analyses that is frequently encountered in the biological literature.

Course content is as follows

Day 1 Course introduction
•   Techniques for data manipulation, aggregation, and visualisation; introduction to linear regression. Packages: {tidyr}, {dplyr}, {ggplot2}

Day 2 Linear models
•   Diagnostics, collinearity, scaling, plotting fitted values); fitting and interpreting interaction terms; model selection and simplification; general linear models and ANCOVA.
•   Packages: {stats}, {car}

Day 3 Generalized linear models
•   Logistic and Poisson regression; predicting using model objects and visualizing model fits.
•   Packages: {broom}, {visreg}, {ggplot2}

Day 4 Mixed effects models
•   Theory and practice of mixed effect models; visualising fixed and random effects.
•   Packages: {lme4}, {broom}, {ggplot2}, {sjPlot}

Day 5 Fitting nonlinear functions
•   Polynomial & Mechanistic models; brief introduction to more advanced topics & combining methods (e.g., generalised linear mixed effects, nonlinear mixed effects, and zero-inflated and zero-altered models).
•   Packages: {nlsTools}.
•   Afternoon to discuss own data if time permits

0
Your rating: None