13  Data Viz

library(tidyverse)

13.1 Axis Breaks

13.1.1 {ggbreak}

df <- data.frame(a = c(1,2,3,500), b = c('a1', 'a2','a3', 'a4'))

ggplot(df) + 
  aes(x = b, y = a) +
  geom_col() +
  ggbreak::scale_y_break(c(5, 490))

Source: {ggbreak} | Set Axis Break for ggplot2

13.1.2 {ggforce}::facet_zoom

df <- data.frame(a = c(1,2,3,500), b = c('a1', 'a2','a3', 'a4'))

ggplot(df) + 
  aes(x = b, y = a) +
  geom_col() +
  ggforce::facet_zoom(ylim = c(0, 5))

Source: Stack Overflow, Sep 2011 | Using ggplot2, can I insert a break in the axis?

13.2 Annotation outside panel

The one time I’ve done this is to add a axis break symbol for when the y axis does not start at zero. In addition to adding the symbol, you also have to set the ggplotGrob to not clip items outside the plot panel.

base <- mpg %>% 
  ggplot(aes(x = cty, y = hwy)) + 
  geom_point() + 
  lims(x = c(0, NA)) + 
  theme_minimal() + 
  theme(axis.line = element_line(color = "grey35", size = 0.25))
Warning: The `size` argument of `element_line()` is deprecated as of ggplot2 3.4.0.
ℹ Please use the `linewidth` argument instead.
p.zoomin <- base + 
  annotate("text"
    , x = -Inf, y = -Inf #bottom left corner of plot 
    , label = "\u2e17"   #symbol ⸗ (diag double hyphen)
    , color = "grey35"   #same color as axis.line
    , size = 4.5         #adjust based on out plot size
    , vjust = -0.25      #adjust further up 
  ) 

#get grob instructions 
g.zoomin <- ggplotGrob(p.zoomin) 

#turn panel clipping off 
g.zoomin$layout$clip[g.zoomin$layout$name == "panel"] = "off"

#draw final 
cowplot::ggdraw(
  cowplot::add_sub(
    g.zoomin
    , "'broken axis' symbol, also extra note under plot"
    , size = 12
    , x = 0
    , hjust = 0
    #, y = 1.25 #same line as x axis title 
  ))

Sources: Stack Overflow, Mar 2014 | set layout clip to ‘off’ and {cowplot}::add_sub | add annotation under plot

13.3 Secondary Legend

13.3.1 NA Legend

Added automatically if NA is part of discrete legend, but is NOT included automatically for continuous legend.

set.seed(20220713)
states_sf <- urbnmapr::get_urbn_map(map = "states", sf = TRUE) %>% 
  #set new CRS 
  sf::st_transform(states_sf, crs = 4326) %>% 
  mutate(
      val_dis = sample(c("A", "B", "C"), nrow(.), TRUE)
    , val_con = sample(-50:50          , nrow(.), TRUE)
  ) %>% 
  mutate(
      val_dis = ifelse(state_abbv == "MN", NA, val_dis)
    , val_con = ifelse(state_abbv == "MN", NA, val_con)
  )
old-style crs object detected; please recreate object with a recent sf::st_crs()
ggplot(states_sf) + geom_sf(aes(fill = val_dis))

ggplot(states_sf) + geom_sf(aes(fill = val_con))

Add NA legend to continuous plot:

ggplot(states_sf) + 
  geom_sf(aes(fill = val_con, color = "")) + 
  scale_fill_continuous(na.value = "grey40") + 
  scale_color_manual(values = NA) + 
  guides(
      color = guide_legend("NA", override.aes = list(fill = "grey40"))
    , fill = guide_colorbar(order = 1)
  )

Source: Stack Overflow, Mar 2017 | Add a box for NA values for continuous map

13.3.2 2+ Special Values

library(cowplot) 

Attaching package: 'cowplot'
The following object is masked from 'package:lubridate':

    stamp
#special value for 'suppressed' data
states_sf <- states_sf %>% 
  mutate(val_con = ifelse(state_abbv == "CA", 999, val_con))

## Main Plot 
p.main <- ggplot(states_sf) + 
  geom_sf( # non-special values, typical cont. color bar 
    data = filter(states_sf, val_con != 999 & !is.na(val_con))
    , aes(fill = val_con), color = NA
  ) + 
  geom_sf( # NA states (missing data) 
    data = filter(states_sf, is.na(val_con))
    , fill = "grey30", color = NA
  ) + 
  geom_sf( # Special values (suppressed data) 
    data = filter(states_sf, val_con == 999)
    , fill = "grey80", color = NA
  ) + 
  geom_sf( # add borders for all 
    fill = NA, color = "white"
  ) 

## Dummy Plot for 2nd Legend 
p.missup <- tibble(val = c(0, 1), type = c("Missing", "Suppressed")) %>% 
  ggplot() + 
  geom_bar(aes(val, fill = type)) + 
  scale_fill_manual(name = NULL
    , values = c("grey30", "grey80")
    , guide = guide_legend( byrow = TRUE)
  ) + 
  theme(legend.spacing.y = unit(5, "pt"))

## Pull out legends and combine 
leg.main <- cowplot::get_legend(p.main)
leg.sec  <- cowplot::get_legend(p.missup)
leg.final <- cowplot::plot_grid(leg.main, leg.sec, ncol = 1, align = "v") 

## Remvoe main legend from plot 
p.final  <- p.main  + theme(legend.position = "none") 

## Combined everything 
cowplot::plot_grid(p.final, leg.final, nrow = 1, rel_widths = c(4, 1))

Quick Link: Stack Overflow, Sep 2018 | Use cowplot for custom multiple legends arrangement

OUTSTANDING

  • how to minimize space between main legend and secondary legend
  • grid::grid.draw(rbind(leg.main, leg.sec)) binds together without huge space, but need to figure out how to add to another ggplot item

13.4 hjust/vjust

Why do I always forget the direction of these?

hjust: 0 = left-aligned, 0.5=center, 1 = right-aligned
vjust: 0 = top-aligned, 0.5=middle, 1 = bottom-aligned

hjustvjust <- expand.grid(
  hjust=c(0, 0.5, 1),
  vjust=c(0, 0.5, 1),
  angle=c(0, 45, 90),
  text="TEXT")

ggplot(hjustvjust, aes(x=hjust, y=vjust)) + 
  geom_point() +
  geom_text(aes(label=text, angle=angle, hjust=hjust, vjust=vjust)) + 
  facet_grid(~angle) +
  scale_x_continuous(breaks=c(0, 0.5, 1), expand=c(0, 0.2)) +
  scale_y_continuous(breaks=c(0, 0.5, 1), expand=c(0, 0.2)) + 
  labs(
    title="Various values of angle, hjust and vjust in ggplot \n"
    , y ="vjust \n"
    , x =" \n hjust"
  ) +  
  theme_bw() + 
  theme(
    axis.text  = element_text(color="black", face="bold" , size=12)
  , axis.title = element_text(color="black", face="plain", size=24)
  , plot.title = element_text(color="black", face="bold" , size=18)
  )

Sources: R bloggers, Nov 2012 | hjust and vjust and GitHub: cdesante/hjustvjust.r

13.5 Math Expressions

13.5.1 Use quote()

ggplot(mpg, aes(displ, hwy))+geom_point()+
  ggtitle(
    quote(
      alpha ^ 2 - frac(1, 10) + sum(n[i], i==1, N)
    ) #end quote
  ) #end of ggtitle

13.5.2 Use TeX() from the latex2exp package

  • must be in a string
  • must be denoted as math mode with dollar signs
  • must include 2 backslashes for \(\LaTeX\) commands
library(latex2exp)
ggplot(mpg, aes(displ, hwy))+geom_point()+
  ggtitle(TeX(
    "$\\alpha^2 - \\frac{1}{10} + \\sum_{i}^N n_i$"
    ) #end TeX
  ) #end ggtitle

13.6 Facets: scales/space

Another option is facet_wrap() or facet_grid(), which can works if the axes are the same for the different variables you want to compare, but be careful as facets are supposed to be comparing items with the same measurements.

tidy.df <- tidyr::pivot_longer(
  mpg, c(class, drv)
  , names_to = "category"
  , values_to = "type"
)

scales = "free": removes types from the axis if that category has 0 cars of that type

ggplot(tidy.df, aes(type)) +
  geom_bar() + coord_flip() + 
  facet_wrap(~category,  ncol = 1, scales = "fixed") + 
  labs(title = 'facet_wrap: scales = "fixed"')
ggplot(tidy.df, aes(type)) +
  geom_bar() + coord_flip() + 
  facet_wrap(~category,  ncol = 1, scales = "free") + 
  labs(title = 'facet_wrap: scales = "free"')

space = "free": spaces based on number of obs (i.e. number of bars); rather than giving each facet equal sizing, ONLY available for facet_grid

ggplot(tidy.df, aes(type)) +
  geom_bar() + coord_flip() + 
  facet_grid(category ~ ., scales = "fixed" , space  = "fixed") + 
  labs(title = 'facet_grid: scales = "fixed" and space = "fixed"')
ggplot(tidy.df, aes(type)) +
  geom_bar() + coord_flip() + 
  facet_grid(category ~ ., scales = "free" , space  = "free") + 
  labs(title = 'facet_grid: scales = "free" and space = "free"')

13.7 Align Axes

Sometimes I’m working on two different types of plots (like a bar chart and a scatter plot) that happen to have the same x-axis. I want to line up these axes so that when the plots are stacked the values correspond to the same date.

13.7.1 gridExtra::grid.arrange() and cowplot::plot_grid()

# two different bar charts
A <- ggplot(mpg, aes(class))+geom_bar()+coord_flip()+ylim(0, 109)
B <- ggplot(mpg, aes(drv))+geom_bar()+coord_flip()+ylim(0, 109)

Using grid.arrange command from the gridExtra package does not line up axes.

#axes don't line up
gridExtra::grid.arrange(A, B, ncol=1)

Use grid.draw command from the grid package to left align graph edges .

Source: Stack Overflow, Nov 2012 | left align two graph edges

The cowplot::plot_grid() function allows you to line up plots by a specific axis.

#make plots into Grobs (grid graphical object)
gA <- ggplotGrob(A) 
gB <- ggplotGrob(B)
grid::grid.draw(rbind(gA, gB))
cowplot::plot_grid(A, B, ncol = 1, align = "v")

13.7.2 Mixed Geoms

Scatter plots and bar charts will not line up automatically, even when using the grid.draw/plot_grid command detailed above. This is because their default limits are different given that the bar chart is centered on the value and the scatter plot is a single point on the value.

#work with smaller subset of economics (ggplot2)
startdate <- "2014-06-01"
economics_small <- economics %>%
  filter(date >= as.Date(startdate)) %>%
  arrange(date)
A <- ggplot(economics_small, aes(date, unemploy))+
  geom_bar(stat="identity")+
  geom_vline(xintercept = as.Date(startdate), color="red", size=2)
Warning: Using `size` aesthetic for lines was deprecated in ggplot2 3.4.0.
ℹ Please use `linewidth` instead.
B <- ggplot(economics_small, aes(date, uempmed))+
  geom_point()+geom_line()+
  geom_vline(xintercept = as.Date(startdate), color="red", size=2)
gA <- ggplotGrob(A) 
gB <- ggplotGrob(B)
#grid::grid.draw(rbind(gA, gB)) 
cowplot::plot_grid(A, B, ncol = 1, align = "v")

In order to line the up there a a couple of options.

13.7.2.1 Fix xlim

If you make the limit the first x-value, the bar chart will not show up (remember it’s centered over the value).

A <- ggplot(economics_small, aes(date, unemploy))+
  geom_bar(stat="identity")+
  geom_vline(xintercept = as.Date(startdate), color="red", size=2)+
  xlim(as.Date(startdate), NA)
B <- ggplot(economics_small, aes(date, uempmed))+
  geom_point()+geom_line()+
  geom_vline(xintercept = as.Date(startdate), color="red", size=2)+
  xlim(as.Date(startdate), NA)
gA <- ggplotGrob(A) 
Warning: Removed 1 rows containing missing values (`geom_bar()`).
gB <- ggplotGrob(B)
#grid::grid.draw(rbind(gA, gB))
cowplot::plot_grid(A, B, ncol = 1, align = "v")
Warning: Removed 1 rows containing missing values (`geom_bar()`).

This can be fixed by adding a half unit to the x-axis (i.e. having the lower limit be half-unit lower than smallest x-value). In this case the unit is a month, so a half-unit would be ~15 days.

HalfUnit <- .5*(economics_small$date[2] - economics_small$date[1])
HalfUnit
Time difference of 15 days
A <- ggplot(economics_small, aes(date, unemploy))+
  geom_bar(stat="identity")+
  geom_vline(xintercept = as.Date(startdate), color="red", size=2)+
  xlim(as.Date(startdate)-HalfUnit, NA)
B <- ggplot(economics_small, aes(date, uempmed))+
  geom_point()+geom_line()+
  geom_vline(xintercept = as.Date(startdate), color="red", size=2)+
  xlim(as.Date(startdate)-HalfUnit, NA)
gA <- ggplotGrob(A) 
gB <- ggplotGrob(B)
#grid::grid.draw(rbind(gA, gB))
cowplot::plot_grid(A, B, ncol = 1, align = "v")

13.7.2.2 Shift Bar Chart

Bar charts are automatically centered over the x-value. Bar charts (and any geom object) can be shifted by using position - position_nudge()). The shift needs to be half a unit on the x-axis, again here it is monthly data so a half unit would be ~15 days.

Source: Stack Overflow, Aug 2016 | shift geom bar right

A <- ggplot(economics_small, aes(date, unemploy))+
  geom_bar(stat="identity", position = position_nudge(x = as.vector(HalfUnit)))+
  geom_vline(xintercept = as.Date(startdate), color="red", size=2)
B <- ggplot(economics_small, aes(date, uempmed))+
  geom_point()+geom_line()+
  geom_vline(xintercept = as.Date(startdate), color="red", size=2)
gA <- ggplotGrob(A) 
gB <- ggplotGrob(B)
#grid::grid.draw(rbind(gA, gB))
cowplot::plot_grid(A, B, ncol = 1, align = "v")

13.8 Default ggplot2

13.8.1 Colors

Default colors can be recreated using scales::hue_pal() and can be shown using scales::show_col()

scales::show_col(scales::hue_pal()(4))

Source: Statistics Globe | Extract Default Color Palette of ggplot2 R Package - Hex Codes

13.9 Geom Defaults

Look up geom’s default aes using: GeomName$default_aes$

Example: geom_text()

GeomText$default_aes
Aesthetic mapping: 
* `colour`     -> "black"
* `size`       -> 3.88
* `angle`      -> 0
* `hjust`      -> 0.5
* `vjust`      -> 0.5
* `alpha`      -> NA
* `family`     -> ""
* `fontface`   -> 1
* `lineheight` -> 1.2