Course: "Getting started with R" https://www.physalia-courses.org/courses-workshops/course13/
18-22 June 2018, Berlin (Germany) Application deadline is: May 20th, 2018. Instructor: Dr. Alexandre Courtiol (Leibniz Institute for Zoo and Wildlife Research, Berlin (Germany)) https://www.physalia-courses.org/instructors/t40/ Session content Monday 18th – Classes from 09:30 to 17:30 Monday is DATA day This first day will be dedicated to the data. R is software dedicated to data analysis, so mastering the basics of data manipulation in R is essential for further steps. It will be explained how to import data into R and how to manipulate them (e.g. from adding or removing rows or columns, to merging tables and using pivot tables). This will be good practice for students to learn the basics of the R language. We will illustrate how to do everything using R base (that is R out of the box), but we will also introduce that allow users to perform some operations on large datasets a little faster. Tuesday 19th – Classes from 09:30 to 17:30 Tuesday is PLOTTING day Plotting is a crucial part of any data analysis no matter how advanced you are in statistics. It is important to visualise the data before the analysis (e.g. to visually check the presence of potential errors and for getting a sense of the distribution of the data), during the analysis (e.g. to check the distribution of model residuals), and after the analysis (to communicate findings in the most efficient way). Therefore, knowing how to plot various kind of data matters a lot. We will thus show how to plot different types of data in R (points, distributions, rasters...) using different graphic environments (e.g graphics, lattice, ggplot2). Wednesday 20th – Classes from 09:30 to 17:30 Wednesday is FUNCTIONS day As John Chambers -- the grandfather of R -- put it "Everything that happens in R is a function call". That R allows for so-called functional programming is one of its great benefits: it allows the implementation of any workflow of statistical analysis as the succession of simple clearly identified steps. Each step is described by a function that takes an input and generates an output. The output of one function is often the input of the next. On this third day we will show that creating one's own function is very very easy (yes, even YOU can program!) and can be very very useful. As a first application we will create our own functions to implement a randomisation test. As a second application we will show how to create functions to perform a power analysis (the estimation of the probability of getting true positives when applying a test) for any statistical test. Thursday 21st – Classes from 09:30 to 17:30 Thursday is LM day Many widely used statistical methods (t-tests, anova, ancova, linear regressions...) are just different type of Linear Models (LM), which is why LM represent the most useful statistical toolbox to be familiar with if you are in natural or social sciences. Fitting a LM in R is easy -- but building models, checking model assumptions, interpreting the outputs, and plotting predictions correctly requires some know how. After refreshing the most important concepts surrounding LM, we will go through each of these steps in detail, working on real datasets. It is important to understand LM quite well before jumping into the more complex methods which we will see on the next day. Friday 22nd – Classes from 09:30 to 17:30 Friday is GLMM day On this last day we will continue with LM and explore what to do when everything goes wrong; starting with when the process generating the data is not gaussian, when the observations are not independent. We will see that the solution is often to use Generalised Linear Models (GLM) which can handle binary and count data and to use Linear Mixed-effects Models (LMM) which can handle non-independence in the data. Both can also be combined under Generalised Linear Mixed-effects Models (GLMM). We will show some specific application of such models (e.g. how LMM can be used to study heredity). We will then see how to tackle spatial and temporal sources of independence in the data, non constant residual variance, and other LM-related pathologies. We will illustrate (G)LM(M) mainly using two R packages: lme4 and spaMM. For more information about the course, please visit our website: https://www.physalia-courses.org/courses-workshops/course20/ Here is the full list of our courses and Workshops: https://www.physalia-courses.org/courses-workshops/ -- MORPHMET may be accessed via its webpage at http://www.morphometrics.org --- You received this message because you are subscribed to the Google Groups "MORPHMET" group. To unsubscribe from this group and stop receiving emails from it, send an email to morphmet+unsubscr...@morphometrics.org.