Submitted to: Field Crops Research
Publication Type: Review Article
Publication Acceptance Date: 5/17/2013
Publication Date: 5/23/2013
Citation: Thorp, K.R. 2013. A book review of Spatial data analysis in ecology and agriculture using R. Field Crops Research. 149:261. Interpretive Summary:
Technical Abstract: Spatial Data Analysis in Ecology and Agriculture Using R is a valuable resource to assist agricultural and ecological researchers with spatial data analyses using the R statistical software(www.r-project.org). Special emphasis is on spatial data sets; how-ever, the text also provides ample guidance on the use of R for datasets that are non-spatial. The text has many strengths. Foremost, it is useful as a statistical reference manual for research applications in agriculture and ecology. Most of the common statistical approaches used by researchers in these fields are reviewed, summarized, and demonstrated in the text. The demonstrations are based on four spatial data sets: two with an ecological theme and two with emphasis in agriculture. A second strength is that the text thoroughly demonstrates the use of the R software, which is a relatively novel open-source statistical package resulting from a collaborative effort among numerous researchers and statisticians worldwide. The text assumes no prior experience with R, but is equally useful for both the novice and the expert. Many R coding examples are included, which demonstrate the implementation of common statistical approaches to analyze the agricultural and ecological data sets provided. A final strength is that the text addresses the spatial nature of these data sets. Discussion of the issues associated with spatial data analysis and the statistical methods for handling spatial data, along with R coding examples, provide insights for using R to process and analyze spatial data sets. Given these strengths, the text provides a handy reference manual for R statistical analyses in the agricultural and ecological sciences. One concern with the text is its organization and layout, as many chapters are presented in a somewhat illogical order. For example, analysis of data from controlled experiments (Chapter 16) should likely be a precursor to mixed models (Chapter 12). However, mixed models are presented several chapters earlier, and the two chapters are not well related. Similarly, the simple topic of multiple regression (Chapter 9) is presented after the more complex topic of multivariate analysis (Chapter 8). Another criticism involves the presentation style for data analysis examples. Often, an example is presented without first describing its goal or purpose, and the reader is led through many data analysis steps with unforeseen twists and turns and no clear objective. These examples can some-times lead to inconclusive results, which can be frustrating for a reader aiming to quickly understand a particular approach by following a concise, meaningful example. An important advantage of open-source software is increased software transparency, since users are free to study, modify, and potentially improve the software’s source code. The opportunity afforded to research scientists is particularly great, because they no longer have to rely on “black box” algorithms for data processing and statistical analysis. With active user and developer communities worldwide, R will continue to be an important statistical software package for the agricultural and ecological sciences. In support of that effort, Spatial Data Analysis in Ecology and Agriculture Using R is a valuable guidebook for using R to conduct statistical analyses on the spatial data sets typically encountered in these sciences.