Chapter 1 Preface

I wrote this online book to accompany the course BB852 - Data Handling, Visualisation and Statistics.

Note: The book is a “work in progress” and will change during the course. The latest version can always be found at the website.

Please let me know (jones@biology.sdu.dk) if you spot any errors, omissions or if you have suggestions for improvement.

The book is divided into three parts: data wrangling, data visualisation and statistics.

1.1 Data wrangling

The term data wrangling covers manipulation of data, for example collected from an experiment or observational study, from its raw form to a form that is ready for analysis, or summarised into tables. It includes reshaping, transforming, filtering and augmenting from other data. This book covers these processes in R mainly using the tools from the dplyr package.

1.2 Data visualisation

Graphing data is a crucial analytical step that can both highlight problems with the data (e.g. errors and outliers) and can inform on appropriate statistical approaches to take. This book covers the use of ggplot2 to make high quality, publication-ready plots.

1.3 Statistics

Statistics is a huge field; I don’t attempt to cover more than a small fraction of it with this course. Instead, I focus on (ordinary) linear models and generalised linear models. In a nutshell, linear models model the effects of explanatory variables on a continuous response variable with a Gaussian (normal) error distribution. In contrast, generalised linear models (GLMs) offer a more flexible approach that allows the response variable to have non-normal error distributions. This flexibility enables the more-appropriate modelling of phenomena, including integer counts (e.g. the number of individuals, species, or events), binary (0/1) data (e.g. survived/died) or binomial count data (e.g. counts of successes and failures). It is essential to realise that most commonly-used statistical methods, including t-tests, ANOVA, ANCOVA, n-way ANOVA, and linear and multiple regression, are all special cases of linear models.

My general approach with communicating these methods and ideas is to teach using examples. Therefore, the bulk of the text here consists of walk-throughs of manipulating, plotting and analysing real data. For the statistics section I focus on communicating the “gist” of the underlying mathematical machinery rather than the mathematical details. If you find yourself interested in these details then there are more specialist textbooks available.

This book accompanies the course lectures. The general idea is that there will be a lecture, followed by computer work where you work through the examples in the relevant chapter of this book. At the end of most chapters there are also exercises to test your new skills. It is very important that you do these to gradually build up your skill level and confidence.

1.4 Data sources

This book uses numerous data sets in examples, most of which are real data sets obtained from published works, or collected by me.

The data sets can be found at the following link: https://www.dropbox.com/scl/fo/5tdl9dtflv79lkvq86vuj/h?rlkey=spw81m08re1ufef5uvxcopgla&dl=0

1.5 Your instructor(s)

You are welcome to contact instructors with any problems/questions (but please put a little effort in first).

1.6 Expectations

There are lectures and practical exercise sessions on the course. The exercise sessions are essential to understand the subject and I expect students to attend and actively participate in them. I also expect students to make every effort to keep up with the core reading (mainly the textbook chapters), and to ask questions where they don’t understand. The nature of the course is that it builds sequentially and therefore, if you miss classes or fall behind, it may be hard to catch up.

If you don’t finish off exercises in class time you should finish them as homework.

1.7 Your feedback

I aim to make this course useful and rewarding for you. I would really like your feedback on how the course is progressing so I can address any issues that come up as soon as possible. To help with this I have created a simple Google Form: ​https://forms.gle/wZhUfy35ZxEmomYt6​. You can use this to send me (Owen) comments (anonymously if you wish) at any time in the course. I promise to do my best to resolve any problems.

1.8 Assessment

Your final grade will be determined by a multiple choice question (MCQ) exam, worth 30% of your grade, and a written assignment, worth 70% of your grade. The take-home assignment will include several questions where you apply your new skills to analyse some interesting data and report your findings.

1.9 Acknowledgements

These materials are inspired by the excellent textbook, “Getting Started With R”1, which is the recommended textbook for BB852, and by materials for the Sheffield University course “AP 240 - Data Analysis And Statistics With R” (https://dzchilds.github.io/stats-for-bio/). For your convenience, the data sets for the Beckerman et al. book are available at this course’s data Dropbox link (above), in a folder called “GSWR_datasets”.

  1. Beckerman, Childs & Petchey (2017) Getting Started With R. Oxford University Press (2nd edition)↩︎