Get familiarized with metadata - Acacia drepanolobium Surveys

ggplot

• Very popular plotting package
• Good plots quickly
• Declarative - describe what you want not how to build it
• Constrasts w/Imperative - how to build it step by step

Data

• Data on acacia size in an experiment in Africa excluding large herbivores
• Data is tab separated
• Includes information on if the plant is dead in the HEIGHT column

acacia <- read.csv("http://www.esapubs.org/archive/ecol/E095/064/ACACIA_DREPANOLOBIUM_SURVEY.txt", sep="\t", na.strings = c("dead"))

Basics

library(ggplot2)

• ggplot() arguments:
  • default dataset - what data are we working with
  • set of mappings
    • ‘Aesthetics’ from variables
    • what columns should we use for different aspects of the plot
  • ggplot(data = acacia, mapping = aes(x = CIRC, y = HEIGHT))
• Add components of figures with layers
  • geom_point()
• Scatter plot showing branch circumference and height

ggplot(acacia, aes(x = CIRC, y = HEIGHT)) +
  geom_point()

• To change things about the layer pass arguments to the geom

ggplot(acacia, aes(x = CIRC, y = HEIGHT)) +
  geom_point(size = 3, color = "blue", alpha = 0.5)

• Rescale axes
  • scale_continuous()

ggplot(acacia, aes(x = CIRC, y = HEIGHT)) +
  geom_point(size = 3, color = "blue", alpha = 0.5) +
  scale_y_log10() +
  scale_x_log10()

• Not changing the data itself, just the presentation of it

• Add Labels (documentation for your graphs!)

ggplot(acacia, aes(x = CIRC, y = HEIGHT)) +
  geom_point(size = 3, color = "blue", alpha = 0.5) +
  labs(x = "Circumference [cm]", y = "Height [m]",
       title = "Acacia Survey at UHURU")

Do Tasks 1-2 in Mass vs Metabolism.

Grouping

• Group on a single graph
• Look at influence of experimental treatment

ggplot(acacia, aes(x = CIRC, y = HEIGHT, color = TREATMENT)) +
  geom_point(size = 3, alpha = 0.5)

• Facet specification

ggplot(acacia, aes(x = CIRC, y = HEIGHT)) +
  geom_point(size = 3, alpha = 0.5) +
  facet_wrap(~TREATMENT)

• Where are all the acacia in the open plots? (eaten?)

Do Tasks 3-4 in Mass vs Metabolism.

Assign Tasks 1-4 in Adult vs Newborn Size.

Layers

• We’ve seen that ggplot makes graphs by combining information on
  • Data
  • Mapping of parts of that data to aspects of the plot
  • A geometric object to represent the data

ggplot(acacia, aes(x = CIRC, y = HEIGHT)) +
  geom_point()

• Many kinds of geometric object (type geom_ and show completions)

• Usage
  • ggplot() sets defaults for layers
  • Can combine multiple layers using +
    • Order matters
• Combine different kinds of layers
• Add a linear model

ggplot(acacia, aes(x = CIRC, y = HEIGHT)) +
  geom_point() +
  geom_smooth(method = "lm")

• Do this by treatment

ggplot(acacia, aes(x = CIRC, y = HEIGHT, color = TREATMENT)) +
  geom_point() +
  geom_smooth(method = "lm")

Do Task 5 in Adult vs Newborn Size.

Statistical transformations

• Geoms include statistical transformations
• So far we’ve seen
  • identity: the raw form of the data or no transformation
  • smooth: model line (e.g., loess, lm)
• Transformations also exist to make things like histograms, bar plots, etc.

• Occur as defaults in associated Geoms

• To look at the number of acacia in each treatment use a bar plot
  • geom_bar()

ggplot(acacia, aes(x = TREATMENT)) +
  geom_bar()

• Uses the transformation stat_count()
  • Counts the number of rows for each treatment
• To look at the distribution of circumferences in the dataset use a histogram
  • geom_histogram()

ggplot(acacia, aes(x = CIRC)) +
  geom_histogram()

• Uses stat_bins() for data transformation
  • Splits circumferences into bins and counts rows in each bin
• Uses bins argument to split data into groups
  • Defaults to bins = 30
• These can be combined with all of the other ggplot2 features we’ve learned

ggplot(acacia, aes(x = CIRC)) +
  geom_histogram(bins = 15) +
  scale_x_log10() +
  facet_wrap(~TREATMENT) +
  labs(x = "Circumference", y = "Number of Individuals")

Do Tasks 1-2 in Sexual Dimorphism Exploration.

Combining different data and aesthetics

• Add tree size data for context
• Layers are plotted in the order they are added

trees <- read.csv("http://www.esapubs.org/archive/ecol/E095/064/TREE_SURVEYS.txt",
                  sep="\t", na.strings = c("dead", "missing", "MISSING", "NA"))
ggplot() +
  geom_point(data = trees, aes(x = CIRC, y = HEIGHT), color = "gray") +
  geom_point(data = acacia, aes(x = CIRC, y = HEIGHT), color = "red") +
  labs(x = "Circumference [cm]", y = "Height [m]")

• Each layer will default to ggplot() mappings unless modified
  • So, we don’t have to specify the arguments that are the same

ggplot(mapping = aes(x = CIRC, y = HEIGHT)) +
  geom_point(data = trees, color = "gray") +
  geom_point(data = acacia, color = "red") +
  labs(x = "Circumference [cm]", y = "Height [m]")

Do Task 3 in Sexual Dimorphism Exploration.

Grammar of graphics

• Geometric object(s)
  • Data
  • Mapping
  • Statistical transformation
  • Position
• Coordinates
• Facets
• In combination uniquely describes any plot

Saving plots as new files

ggsave(“acacia_by_treatment.jpg”)

• Lots of optional arguments
  • Location
  • Type
  • Size

ggsave(“figures/acacia_by_treatment.pdf”, height = 5, width = 5)

Assign the rest of the exercises.