R for Data Science¶
The contents here is based on R for Data Science. Check chapter 3, Data Visualization and chapter 5, Data Manipulation for more details.
Here, I'd like to cover
- Data Visualization using ggplot
- Data Manipulation using dplyr
Data Visualization¶
First call tidyverse library
In [1]:
library(tidyverse)
We will play with mpgdata containing observations collected by the US Environmental Protection Agency on 38 models of car.
In [2]:
mpg[1:5,]
Aesthetic Mappings¶
ggplot(data = data, mapping = aes(x = , y=, color= , size= , alpha= , shape=) )
x and y¶
In [4]:
ggplot(data = mpg, mapping = aes(x = displ, y = hwy) ) +
geom_point()
color¶
In [5]:
ggplot(data = mpg, mapping = aes(x = displ, y = hwy, color = class) ) +
geom_point()
size¶
In [6]:
ggplot(data = mpg, mapping = aes(x = displ, y = hwy, size = cyl) ) +
geom_point()
alpha¶
In [7]:
ggplot(data = mpg, mapping = aes(x = displ, y = hwy, alpha = cyl) ) +
geom_point()
shape¶
In [11]:
ggplot(data = mpg, mapping = aes(x = displ, y = hwy, shape = factor(cyl)) ) +
geom_point()
Facets¶
Split your plot into facets (facet_wrap, facet_grid)
- facet_wrap for 1d split
- facet_grid for 2d split
facet_wrap¶
In [12]:
ggplot(data = mpg, aes(x = displ, y = hwy)) +
geom_point() +
facet_wrap(~ class, nrow = 2)
facet_grid¶
In [13]:
ggplot(data = mpg, aes(x = displ, y = hwy)) +
geom_point() +
facet_grid(drv ~ cyl)
Geometric objects¶
Functions start with geom_
- Plot geometrical objects
In [14]:
ggplot(data = mpg, aes(x = displ, y = hwy)) +
geom_point()
In [16]:
ggplot(data = mpg, aes(x = displ, y = hwy)) +
geom_point() +
geom_smooth()
geom_bar¶
In [17]:
ggplot(data = mpg, aes(x = cyl) ) +
geom_bar()
In [19]:
demo <- tribble(
~cut, ~freq,
"Fair", 1610,
"Good", 4906,
"Very Good", 12082,
"Premium", 13791,
"Ideal", 21551
)
demo
In [20]:
ggplot(demo, aes(x = cut, y = freq)) +
geom_bar(stat = "identity")
Theme¶
In [21]:
ggplot(demo, aes(x = cut, y = freq)) +
geom_bar(stat = "identity") +
theme_bw()
In [25]:
ggplot(demo, aes(x = cut, y = freq)) +
geom_bar(stat = "identity") +
theme_bw() +
theme(panel.border = element_blank(),
axis.line = element_line(size = .8, color = "black"),
panel.grid.minor = element_blank(),
text = element_text(size = 25)
)
Data manipulation¶
x %>% f(y) turns into f(x,y), and x %>% f(y) %>% g(z) turns into g(f(x,y), z)
pipe operator¶
filter¶
filtering rows
In [27]:
mpg[1:10,]
In [30]:
mpg %>% filter(year == 1999)
In [31]:
mpg %>% filter(year == 1999, model == 'passat')
In [32]:
mpg %>% filter(year == 1999, model %in% c('passat', 'new beetle') )
select¶
Return a subset of the columns
In [33]:
mpg %>%
filter(year == 1999, model %in% c('passat', 'new beetle') ) %>%
select(model, displ, class)
arrange¶
Reorders the rows according to single or multiple variables
In [34]:
mpg %>%
filter(year == 1999, model %in% c('passat', 'new beetle') ) %>%
select(model, displ, class) %>%
arrange(displ)
In [35]:
mpg %>%
filter(year == 1999, model %in% c('passat', 'new beetle') ) %>%
select(model, displ, class) %>%
arrange(desc(displ))
mutate¶
Add columns from existing data
In [39]:
mpg %>%
filter(year == 1999, model %in% c('passat', 'new beetle') ) %>%
select(model, displ, class) %>%
arrange(desc(displ)) %>%
mutate(displ10 = displ*10,
model2 = car::recode(model, "'passat'='p'; 'new beetle'='nb' "),
ex_missing = car::recode(displ, "2.0 = '20'; 2.8 = '28'; else = NA"))
In [48]:
mpg %>%
filter(year == 1999, model %in% c('passat', 'new beetle') ) %>%
select(model, displ, class) %>%
arrange(desc(displ)) %>%
mutate(displ10 = displ*10,
model2 = car::recode(model, "'passat'='p'; 'new beetle'='nb' "),
ex_missing = car::recode(displ, "2.0 = '20'; 2.8 = '28'; else = NA"),
ex_missing = factor(ex_missing, levels = c("28", "20")),
ex_missing = fct_explicit_na(ex_missing)
)
summarize¶
Produce summary statistic for each group (using group_by() )
In [50]:
mpg %>%
filter(year == 1999, model %in% c('passat', 'new beetle') ) %>%
select(model, displ, class) %>%
arrange(desc(displ)) %>%
mutate(displ10 = displ*10,
model2 = car::recode(model, "'passat'='p'; 'new beetle'='nb' "),
ex_missing = car::recode(displ, "2.0 = '20'; 2.8 = '28'; else = NA"),
ex_missing = factor(ex_missing, levels = c("28", "20")),
ex_missing = fct_explicit_na(ex_missing)
) %>%
group_by(model) %>%
summarize( n = n(),
mean_displ = mean(displ),
sd_displ = sd(displ))
ggplot followed by dyplyr manipulation¶
In [59]:
mpg %>%
filter(year == 1999, model %in% c('passat', 'new beetle') ) %>%
select(model, displ, class) %>%
arrange(desc(displ)) %>%
mutate(displ10 = displ*10,
model2 = car::recode(model, "'passat'='p'; 'new beetle'='nb' "),
ex_missing = car::recode(displ, "2.0 = '20'; 2.8 = '28'; else = NA"),
ex_missing = factor(ex_missing, levels = c("28", "20")),
ex_missing = fct_explicit_na(ex_missing)
) %>%
group_by(model) %>%
summarize( n = n(),
mean_displ = mean(displ),
sd_displ = sd(displ)) %>%
ggplot(aes(x = model, y = mean_displ )) +
geom_bar(stat = "identity") +
geom_errorbar(aes(ymin=mean_displ-sd_displ, ymax=mean_displ+sd_displ), width=.2) +
theme_bw() +
theme( text = element_text(size = 25) ) +
xlab('Model') +
ylab('Displ')
In [ ]:
In [ ]: