############################################################
### R BASICS WORKSHOP                                    ###
### PRESENTATION 3.2: MAIN CLASSES OF OBJECTS            ###
###                                                      ###
### Center for Conservation and Sustainable Development  ###
### Missouri Botanical Garden                            ###
### Website: rbasicsworkshop.weebly.com                  ###
### Last modification: 2021-June-06                      ###
############################################################


### C. CLASSES OF OBJECTS ##################################

## IMPORTANT: The class of an object is an important 
## determinant of how it behaves in R. 

## Confusing classes is a common source of problems when 
## programming!

# 1. Vectors: 1 dimension - numeric, character or logical
# 2. Factors: 1 dimension - levels of a categorical variable
# 2. Matrices: 2 dimensions - rows by columns
# 3. Arrays: n dimensions
# 4. Data frames: 2 dimensions - observations by variables
# 5. Lists

# 6. Other classes: e.g. lm, phylo, shapefile, etc.

## IMPORTANT: Each object has ATTRIBUTES defining how it 
## works in R.
## Common attributes include: class, length, dimensions, 
## names, etc.



### A. SOME OBJECTS WE WILL NEED LATER #####################

# a. Let's suppose we have abundances of 5 species measured 
#    in a forest plot in two different years (numeric 
#    values). 
abund.t1 <- c(0, 17, 34, 26, 82)
abund.t2 <- c(1, 17, 31, 27, 91)

# b. Let's suppose, we also have the names of the species 
#    (character data).
spp <- c("I.ynga", "I.edulis", 	"I.macrophylla", 
  "I.punctata", 	"I.alba")

# c. Also, let's suppose we need a variable that indicate if  
#    abundances have increased or not (logical values).
increm <- c(TRUE, FALSE, FALSE, TRUE, TRUE)



### B. VECTORS ############################################ 
# Vectors represent a linear sequence of values.
## IMPORTANT: Each value in the sequence is an ELEMENT of 
## the object.

abund.t1
abund.t2
spp
increm

# CLASS is the main attribute of a vector
class(abund.t1)
class(spp)
class(increm)

# LENGTH is another fundamental attribute of a vector
length(x = abund.t1)
length(x = spp)
length(x = increm)


## IMPORTANT: Vectors can ONLY have one main type of data: 
## numeric, character or logical. NEVER A COMBINATION!

x <- c(1,2,3)
x
class(x)

x2 <- c(1,2,3, "a")
x2
class(x2)


## However, NAs can be introduced with any of these types 
## of data:
x <- c(1,2,NA,3,NA)
x
class(x)



### C. FACTORS #############################################
# Factors also represent a linear sequence of values, but 
# are designed to be LEVELS of a categorical variable.

# For example, suppose we have the phenological status of  
# several individual trees:

pheno <- c("sterile", "sterile", "flowering", "fruiting", 
  "sterile", "flowering", "flowering", "sterile")
pheno
length(pheno)
class(pheno)
mode(pheno)

pheno2 <- factor(x = pheno)
pheno2
length(pheno2)
class(pheno2)
mode(pheno2)

identical(pheno, pheno2)

# An important attribute of factors is their LEVELS
levels(pheno2)


# IMPORTANT: the class of an object commonly affects how  
# functions work with them. For example:

plot(pheno)
plot(pheno2)



### D. MATRICES ############################################
# Matrices are similar to vectors, but they have two 
# dimensions: ROWS AND COLUMNS.

# Suppose we have abundances of 5 species of trees (columns) 
# in 3 different sites (rows):
abund.M <- matrix(data = 1:(3*5), nrow = 3, ncol = 5)
abund.M

class(abund.M)


# Let's use a little more realistic values of abundance
abund.M <- matrix(data = rpois(n = 3*5, lambda = 10), 
  nrow = 3, ncol = 5)

abund.M


# Like vectors, matrices have LENGTH which return the 
# number of elements
length(abund.M)


# In addition, matrices also have DIMENSIONS
dim(abund.M)

nrow(abund.M)
ncol(abund.M)


# In matrices, it is often useful to have NAMES for rows 
# and columns
colnames(abund.M)
rownames(abund.M)

colnames(abund.M) <- spp
rownames(abund.M) <- paste("site_", 1:nrow(abund.M), 
  sep = "")

abund.M


## IMPORTANT: Matrices, like vectors, can have ONLY one type 
## of data: numeric or character or logical. NEVER A 
## COMBINATION.

abund.M2 <- matrix(sample(c(1:10, letters), 3*5), 
  nrow = 3, ncol = 5)
abund.M2

class(abund.M2)
mode(abund.M2)


spp
abund.t1
abund.t2

mixed.matrix <- cbind(spp, abund.t1, abund.t2)
mixed.matrix
mode(mixed.matrix)



### E. ARRAYS ##############################################
# Arrays are similar to matrices, but have MORE THAN TWO 
# DIMENSIONS.

# With three dimensions, it's like having a cube of data. 
# Like multiple matrices of the same dimensions one behind 
# the another (aligned by rows and columns).

abund.A <- array(data = 1:(3*5*2), dim = c(3,5,2))
abund.A

# With a little more realistic values of abundance:
abund.A <- array(data = rpois(3*5*2, 10), dim = c(3,5,2))
abund.A

mode(abund.A)
class(abund.A)

length(abund.A)
dim(abund.A)
dimnames(abund.A)

# We can give names to each element in each dimension
dimnames(abund.A) <- list(paste("site_", 1:3, sep = ""), 
  spp, c("2008", "2013"))

abund.A



### F. DATA FRAMES #########################################
# Data frames organize measurements of variables (columns) 
# on study units (rows). Thus, in data frames EACH COLUMN 
# MAY BE OF a DIFFERENT TYPE (numeric, character, logical 
# or factors).

abund.data <- data.frame(spp, abund.t1, abund.t2, increm)
abund.data

class(abund.data)


# IMPORTANT: matrices and data frames behave differently!
abund.data2 <- as.matrix(abund.data)
abund.data2
class(abund.data2)

dim(abund.data)
dim(abund.data2)

length(abund.data)
length(abund.data2)

mode(abund.data)
mode(abund.data2)

names(abund.data)
names(abund.data2)

plot(abund.data)
plot(abund.data2)



### G. LISTS ###############################################
# Lists can also contain elements of different types or 
# classes. Elements in a list DO NOT need to be of the same  
# length or dimensions.

abund.M
class(abund.M)

pheno
class(pheno)

abund.data
class(abund.data)


random.list <- list(abund.M, pheno, abund.data)

# Each object becomes an element of the list
random.list


# Some attributes of a list:
class(random.list)
length(random.list)
dim(random.list)
names(random.list)

names(random.list) <- c("abund.M", "pheno", "abund.data")
random.list


# Understanding the structure of a list can be very useful:
str(random.list)


## NOTE: Many functions use list-like objects to export 
##       results of analyses:

x <- rnorm(100)
y <- 4 + 5 * x + rnorm(100)

plot(y ~ x)

lm.res <- lm(y ~ x)
lm.res

str(lm.res)



### H. OTHER ###############################################
# Many functions generate objects of particular classes. 
# However, the vast majority are modifications of the basic 
# objects that we revised. For example:

# Linear models generate objects of class *lm*, but have 
# the structure of lists:

class(lm.res)

# However, being of class *lm* makes certain functions 
# handle this object in ways specific to its class.

plot(lm.res)


# Another object class are phylogenetic trees 
# (class *phylo*)
install.packages("ape") # This installs a package 
                        # called *ape*
library(ape) # This opens the package


# The package *ape* contains a phylogenetic tree of bats
data(chiroptera) # This opens the phylogenetic tree named 
                 # *chiroptera*
class(chiroptera)

# Despite being an object of class *phylo*, this object has 
# the structure of a list:
str(chiroptera)


# The function *plot* knows what to do when given a 
# phylogenetic tree
plot(chiroptera, show.tip.label = FALSE)