IPS-003 (2023)

Data Analysis and Design of Experiments

Author

Pedro J. Aphalo

R and programing

R is a language and an environment for data analysis and visualization. It has become the standard for data analysis. In Bioinformatics it “competes” with Python, but R can “talk” with code written in most other programing languages. Course IPS-003 focuses on the principles of data analysis and design of experiments. We use R for exercices and demonstrations. Thus some familiarity with R is needed. So, I am organizing a short series of meetings to introduce R. These are not part of IPS-003, and there is no course code for them, but credits will be accrued and registered under a generic code. The pages below will be used and you will initially use R online through Posit Cloud (formerly RStudio Cloud).

Moderately advanced R learning material is available as a free on-line course at intro2R.

The second edition of my book Learn R: As a Language is about to published and I will provide participants with a copy of the PDF.

IPS-003

I will be teaching data analysis and design of experiments, which are very tightly dependent on each other. Statistics gives theoretical support to data analysis methods, but efficiently extracting information from observations from experiments and surveys is in many ways like detective work or solving puzzles. Modern data analysis makes heavy use of visual data displays (plots, diagrams, garphs). Much of the material for the course is in the pages listed below.

Introduction to Data Visualization
Modified 2023-11-17
abstract An introduction to data visualization using R. The same data are plotted in different ways to demonstrate how different approaches to plotting emphasize different aspects of the data. Plots suitable for different audiences are also created. The source code for all examples is included in the page.
Research as a process
Modified 2023-10-19
abstract Fortuitously, in the last few days I have several times have had to ponder about the roles of hypothesis-based and descriptive approaches to research. Being an agronomist and ecologist rather than a philosopher, the question that worries me is how data analysis differs between these two approaches and how we should interpret the results of the analysis in each case. This page collects some of my thoughts and opinions on the subject at the time of writing.
Design of Experiments
Modified 2023-10-21
abstract An introduction to design and planning of experiemnts. The roles replication, randomization, and range of validity.
Randomization and independent replicates
Modified 2023-11-27
abstract In this page I give a brief presentation on how randomization can be constrained to match the properties of the systems under study without compromising the validity of conclusions. This gives origin to different formal designs for experiments, and corresponding data analysis approaches. I also mention pseudoreplication as a problem that needs to be avoided whenever possible, and otherwise the additional implicit assumptions shown to be true or at least highly plausible.
Theoretical probability distributions
Modified 2023-11-21
abstract Shiny Apps or dashboards for interactive explorartion on the dependence of the t and F distributions on the degrees of freedom.
Linear Models
Modified 2023-11-30
abstract In this page I give a brief presentation on the mechanics of fitting statistical models to observed data using R. I use a lineal model (LM), polynomial regression, as example. I describe the most important values contained in the model fit object returned by function lm() and how to extract them. I rely heavily on diagrams and data plots.
EDA with ‘ggplot2’
Modified 2023-07-27
abstract EDA and ‘ggplot2’ have a common ancestry. EDA as originally proposed by John W. Tukey included data visualization as a key tool. In this page I show examples of some of the data visualizations that are used in EDA, together with the R code I used to create them. I also briefly describe EDA’s position in the scientific research process and mention its origins.
No matching items

Why these pages are not in Moodle? Because I want to make sure they are truly open access, and that they will remain available after I formally retire next year as a Senior Lecturer.