Introduction to quantitative methods with

Chapter 3: Representing data with ggplot2

Introduction

Note

• Exercises associated with this chapter here

• Back to the main page

Introduction

• Data visualization synthesizes data structure
  • Guides further data exploration
  • Conveys a message to the reader

• 2 types of visualizations:
  • Static representations: ggplot2
  • HTML representations: leaflet

Example dataset

library(doremifasol)
library(dplyr)
library(ggplot2)
library(scales)

df <- telechargerDonnees("FILOSOFI_DISP_IRIS", date = 2017) %>%
  as_tibble() %>%
  sample_frac(0.1)

Note

Inspiration for these slides:

• Collaborative documentation utilitR;
• ggplot manual;
• Course by Joseph Larmarange

ggplot2

Initialize a figure associated with a dataset

ggplot(df, aes(x = DISP_MED17, y = DISP_D917))

ggplot2

Add layers (+) with geom_* functions

ggplot(df) +
  geom_point(aes(x = DISP_MED17, y = DISP_D917))

ggplot2

Parameterize layers with aes

ggplot(df) +
  geom_point(aes(x = DISP_MED17, y = DISP_D917, color = DISP_Q117), shape = 3)

Note

Aesthetic control of a geom_ layer is done through:

• aes: variable parameters of the layer linked to a variable;
• outside aes: parameters that apply uniformly to the layer

ggplot2

Add layers (+) with geom_* functions

ggplot(df, aes(x = DISP_MED17, y = DISP_D917)) +
  geom_point(aes(color = DISP_Q117), shape = 3) +
  geom_smooth(color = "red", alpha = 0.7, se = FALSE)

ggplot2

Modify scales with scale_ functions

ggplot(df, aes(x = DISP_MED17, y = DISP_D917)) +
  geom_point(aes(color = DISP_Q117), shape = 3) +
  geom_smooth(color = "red", alpha = 0.7, se = FALSE) +
  scale_x_continuous(labels = unit_format(unit = "k", scale=1e-3)) +
  scale_y_continuous(trans='log', labels = unit_format(unit = "k", scale=1e-3)) +
  scale_color_viridis_c()

ggplot2

Modify scales with scale_ functions

df <- df %>% mutate(quartile = factor(ntile(DISP_Q117, 4)))
ggplot(df, aes(x = DISP_MED17, y = DISP_D917)) +
  geom_point(aes(color = quartile), shape = 3) +
  geom_smooth(color = "red", alpha = 0.7, se = FALSE) +
  scale_x_continuous(labels = unit_format(unit = "k", scale=1e-3)) +
  scale_y_continuous(trans='log', labels = unit_format(unit = "k", scale=1e-3)) +
  scale_color_viridis_d(option = "turbo")

ggplot2

Modify aesthetics, only at the end

p <- ggplot(df, aes(x = DISP_MED17, y = DISP_D917)) +
  geom_point(aes(color = quartile), shape = 3) +
  geom_smooth(color = "red", alpha = 0.7, se = FALSE) +
  scale_x_continuous(labels = unit_format(unit = "k", scale=1e-3)) +
  scale_y_continuous(trans='log', labels = unit_format(unit = "k", scale=1e-3)) +
  scale_color_viridis_d(option = "turbo")

p + theme_bw() +
  labs(x = "Median income", y = "9th decile", color = "Quartile") +
  theme(legend.position = "bottom")