Hint Normality of a vector can be tested by using the Shapiro-Wilk test, available
through function shapiro.test().
## Standardization of all environmental variables
# Center and scale = standardize the variables (z-scores)
env.z <- decostand(env, "standardize")
apply(env.z, 2, mean)
# means = 0
apply(env.z, 2, sd) # standard deviations = 1
# Same standardization using the scale() function (which returns
# a matrix)
env.z <- as.data.frame(scale(env))
2.3 Conclusion
The tools presented in this chapter allow researchers to get a general impression of
their data. Although you will see much more elaborate analyses in the following
chapters, keep in mind that a first exploratory look at the data can tell much about
them. Information about simple parameters and distributions of variables is important to consider and will help in the correct selection of more advanced analyses.
Graphical representations like bubble maps are useful to reveal how the variables are
spatially organized; they may help generate hypotheses about the processes acting
behind the scene. Boxplots and simple statistics may be necessary to reveal unusual
or aberrant values.
EDA is often neglected by scientists who are eager to jump to more sophisticated
analyses. We hope we have convinced you that it should have an important place in
the toolbox of ecologists.
34
2 Exploratory Data Analysis
through function shapiro.test().
## Standardization of all environmental variables
# Center and scale = standardize the variables (z-scores)
env.z <- decostand(env, "standardize")
apply(env.z, 2, mean)
# means = 0
apply(env.z, 2, sd) # standard deviations = 1
# Same standardization using the scale() function (which returns
# a matrix)
env.z <- as.data.frame(scale(env))
2.3 Conclusion
The tools presented in this chapter allow researchers to get a general impression of
their data. Although you will see much more elaborate analyses in the following
chapters, keep in mind that a first exploratory look at the data can tell much about
them. Information about simple parameters and distributions of variables is important to consider and will help in the correct selection of more advanced analyses.
Graphical representations like bubble maps are useful to reveal how the variables are
spatially organized; they may help generate hypotheses about the processes acting
behind the scene. Boxplots and simple statistics may be necessary to reveal unusual
or aberrant values.
EDA is often neglected by scientists who are eager to jump to more sophisticated
analyses. We hope we have convinced you that it should have an important place in
the toolbox of ecologists.
34
2 Exploratory Data Analysis