Gráficos con ggplot2 ( )

1 Graficos usando ggplot2 ( )

Se mostrará los distintos gráficos que pueden realizarse desde el paquete ggplot2.

1.1 geom_area ( )

# Usando el Stat= "count"; haciendo eso resulta conteos
airquality %>% ggplot(aes(x= Temp))+
  geom_area(stat = "count", fill="aquamarine", alpha=0.6)+
  theme_bw()

airquality %>% ggplot(aes(x=Temp))+
  geom_area(stat = "bin",fill="aquamarine",alpha=0.6,
            lty=3, lwd=1.5, color="blue")+
  theme_bw()

1.2 geom_freqpoly ( )

airquality %>% ggplot(aes(x=Temp))+
  geom_freqpoly(color="orange", lwd=1.1, lty=2, alpha=0.6)+
  theme_dark()

1.3 geom_density ( )

airquality %>% ggplot(aes(x=Temp))+
  geom_density(fill="coral", alpha=0.5, color="darkblue",
               lty=4,lwd=1.1)+ theme_classic()

1.4 geom_histogram( )

airquality %>% ggplot(aes(x=Temp))+
  geom_histogram(bins = 20, color="red", fill="gray80",
                 lty=3, lwd=1.1)+theme_bw()

1.5 stat_qq_line y stat_qq( )

ggplot(data=airquality,aes(sample=Temp))+
  stat_qq_line(color="darkblue",
               lwd=1.1)+ stat_qq(color= "red", lwd=3,
                                 alpha=0.2)+ theme_bw()

1.6 geom_bar ( )

ggplot(data=airquality, aes(x=Month))+
  geom_bar(fill="deepskyblue",stat = "count", alpha=0.4)

1.7 geom_label ( )

s <- msleep %>% ggplot(aes(x=sleep_total,y=sleep_rem))

s + geom_label(aes(label=vore, fill=vore), alpha=0.25)+
  geom_point(color="red")

1.8 geom_text ( )

s +geom_text(aes(label=vore))+ geom_point(color="red")

1.9 geom_label_repel ( )

s + geom_label_repel(aes(label=vore, fill=vore), alpha=0.6) +
  geom_point(color="red")

1.10 geom_text_repel ( )

s + geom_text_repel(aes(label=vore, color=vore)) + 
  geom_point(color="red")

1.11 geom_jitter ( )

d <- storms %>% ggplot(aes(x=wind, y=pressure))

d + geom_point(color="blue")

d + geom_jitter(aes(color=category),
                width = 6, height = 6,
                alpha=0.4)

1.12 geom_hex ( )

d + geom_hex(aes(fill=category),
             color="gray40")

1.13 geom_density2d ( )

d + geom_density2d()

1.14 stat_density_2d ( )

d + stat_density_2d(aes(fill=..density..),
                    geom = "raster",
                    contour = FALSE) + 
  scale_fill_distiller(palette = "Spectral",
                       direction = -1)

1.15 stat_boxplot ( )

airquality %>% 
  ggplot()+ 
  stat_boxplot(geom="errorbar",
                         width=0.2,
                         aes(x=factor(Month),
                             y=Ozone,
                             fill=Month)) +
  geom_boxplot(aes(x=factor(Month),
                             y=Ozone),
               fill="deepskyblue")

iris %>% ggplot(aes(x=Species,y=Sepal.Width)) +
   stat_boxplot(geom = "errorbar", width=0.2) +
                  geom_boxplot(aes(fill=Species))

iris %>% ggplot(aes(x=Species, y=Sepal.Width, alpha=0.3)) +
  geom_violin(aes(fill=Species))+ geom_boxplot(aes(color=Species))+
  geom_jitter(aes(color=Species), size=4)+
  theme_bw()

1.16 geom_count ( )

midwest %>% ggplot(aes(x=category,y=state, color=state))+
  geom_point()

midwest %>% ggplot(aes(x=category,y=state, color=state))+
  geom_jitter()

midwest %>% ggplot(aes(x=category,y=state, color=state))+
  geom_count()+ scale_size_area(max_size = 10)

1.17 geom_tile ( )

# Usando una base de datos de práctica:

xde <- openxlsx::read.xlsx("BD2.xlsx")

head(xde, n = 5)

  Var1 Var2     score
1    a    1  6.218171
2    b    1 26.568888
3    c    1  6.421618
4    d    1 20.185120
5    e    1  8.716945

xde %>% ggplot()+
  geom_tile(aes(x=factor(Var2),y=Var1,fill=score))+
  scale_fill_gradient(low = "white",high="darkcyan")

xde %>% ggplot()+
  geom_tile(aes(x=factor(Var2),y=Var1,fill=score))+
  scale_fill_gradient2(low = "blue",high="red",
                       midpoint = 15,
                       mid="whitesmoke")

1.18 geom_smooth ( )

bio <- openxlsx::read.xlsx("Datos_bio.xlsx")

# Regresión lineal simple
bio %>% ggplot(aes(x=carbono,y=humedad1))+
  scale_x_continuous(breaks = seq(0,380000,50000))+
  geom_point(color="darkcyan", size=4, alpha=0.25) + 
  geom_smooth(method = "lm",se=TRUE, fill="aquamarine",
              color="red") + 
  theme_bw()

# Regresión lineal cúbica
bio %>% ggplot(aes(x=carbono,y=humedad1))+
  scale_x_continuous(breaks = seq(0,380000,50000))+
  geom_smooth(method = "lm",
              se=FALSE,formula = y~x+poly(x,2),
              color="red", lwd=1.2, lty=2)+
  geom_point(color="purple", size=3, alpha=0.25) + 
  theme_minimal()

# Regresión generalizada aditiva(no lineal)

bio %>% ggplot(aes(x=carbono,y=humedad1))+
  scale_x_continuous(breaks = seq(0,380000,50000))+
  geom_smooth(method = "gam",
              se=FALSE,color="darkgreen", 
              lwd=1.2, lty=2)+
  geom_point(color="skyblue", size=3, alpha=0.25) + 
  theme_minimal()

bio %>% ggplot(aes(x=carbono,y=humedad1))+
  scale_x_continuous(breaks = seq(0,380000,50000))+
  geom_smooth(method = "loess",
              se=FALSE,color="red", 
              lwd=1.2, lty=2)+
  geom_point(color="coral", size=3, alpha=0.25) + 
  theme_minimal()

# Regresión lineal generalizada lineal(GLM), de tipo binomial

bio %>% ggplot(aes(x=biomasa,y=bin))+
  scale_y_continuous(breaks = seq(0,1,1))+
  geom_smooth(method = "glm",
              se=TRUE,color="red", 
              lwd=1.2, lty=2, fill="gray80",
              method.args=list(family=binomial))+
  geom_point(color="coral", size=3, alpha=0.25)+
  theme_minimal()

1.19 geom_vline(), geom_hline()

P <- bio %>% ggplot(aes(x=carbono,y=humedad1))+
  scale_x_continuous(breaks = seq(0,380000,50000))+
  geom_smooth(method = "loess",
              se=FALSE,color="red", 
              lwd=1.2, lty=2)+
  geom_point(color="coral", size=3, alpha=0.25) + 
  theme_minimal()

# geom_vline()
P + geom_vline(xintercept = 200000, color="skyblue", lwd=0.8)

# geom_hline()
P + geom_hline(yintercept = 600,lwd=0.9, color="green") + 
  geom_text(data=data.frame(x=200000,y=580, lab="humbral"),
            aes(x=x,y=y, label=lab),color="blue")

#geom_abline()
coeff <- lm(humedad1~carbono, data=bio) %>% coef()
P + geom_abline(intercept = coeff[1], slope = coeff[2], color="darkblue", lwd=0.9)

1.20 geom_abline()

#geom_abline()
coeff <- lm(humedad1~carbono, data=bio) %>% coef()
P + geom_abline(intercept = coeff[1], slope = coeff[2], color="deepskyblue", lwd=0.9)

1.21 geom_segment()

P + geom_segment(x=385534.359, y=803.14, xend=1612.323, yend=255.85,
                 color="hotpink",lwd=0.9)

1.22 geom_curve()

# curva
P + geom_curve(x=385534.359,
               y=803.14, xend=1612.323, yend=255.85,
               curvature=0.25,color="aquamarine",
               lwd=0.9)

# flecha
P + geom_curve(x=385534.359, y=803.14,
               xend=1612.323, yend=255.85,
               curvature=0.25,
               arrow = arrow(length = unit(0.8,"cm")),
               color="purple",lwd=0.9)

1.23 scale_x_reverse()

Permite cambiar el sentido del eje x para que los valores del eje incrementen de derecha a izquierda.

P + scale_x_reverse()

1.24 scale_y_reverse()

Permite cambiar el sentido de los valores del eje “y” haciendo que los valores incrementen de desde el extremo más alejado del eje, hacia el origen.

P + scale_y_reverse()

1.25 scale_y_log10()

Transformación de los valores del eje “y”, muy util a la hora de comparar observaciones con valores muy diferentes.

bio %>% ggplot(aes(x= especies, y= DBH))+
  geom_bar(stat = "identity")

# Luego de transformar:
bio %>% ggplot(aes(x= especies, y= DBH))+
  geom_bar(stat = "identity") +
  scale_y_log10()

bio %>% ggplot(aes(x= especies, y= DBH))+
  geom_bar(stat = "identity")+
  scale_y_sqrt()

1.26 ylim() y xlim()

Permite visualizar solo los valores de un intervalo propuesto.

P + xlim(0,10000)+ylim(200,600)

# Programando para que los datos del eje salgan en sentido inverso:
P + ylim(2000,500)

1.27 scale_x_continuous() y scale_y_continuous()

P + scale_x_continuous(breaks = seq(0,450000,30000)) + 
  scale_y_continuous(limits = c(250,800),
                     breaks = seq(100,1000,100),
                     expand = c(0,0,0,0))

#### Expand permite controlar los márgenes de gráfico
#### expand=c(y,x,y2,x2)
bio %>% ggplot(aes(x=especies, y=DBH)) +
  geom_bar(stat = "identity") + 
  scale_y_continuous(expand = c(0,0,0.5,0))

# Coerción de etiquetas del eje x:
P + scale_x_continuous(labels = scales::comma)

P + scale_x_continuous(labels = scales::percent)

P + scale_x_continuous(labels = scales::label_percent(scale=2))

P + scale_x_continuous(labels = scales::dollar)

P + scale_x_continuous(labels = scales::scientific)

# valor absoluto
P + scale_x_continuous(labels = abs)

# Con esta función se puede colocar un segundo eje:
ggplot()+
  geom_bar(data = bio, aes(x=especies, y= DBH),
           stat = "identity",color="green") + 
  geom_bar(data = bio, aes(x=especies, y=humedad1),
           stat = "identity",fill="red", alpha=0.25) +
  scale_y_continuous(sec.axis = sec_axis(~.,name = "humedad1"))

bio %>% ggplot() + 
  geom_smooth(aes(x=carbono,y=humedad1),color="red",fill="orange") +
  geom_point(aes(x=carbono,y=humedad1),color="hotpink") +
  geom_smooth(aes(x=carbono,y=biomasa/10),color="blue",fill="deepskyblue")+
  geom_point(aes(x=carbono,y=biomasa/10), color= "purple") +
  scale_y_continuous(sec.axis = sec_axis(~.*10, name="bio")) +
  theme_bw()

1.28 Gráfico de tornado

set.seed(123)
n1 <- abs(rnorm(10)) + 0.1
prob <- n1/sum(n1)

data1 <- data.frame(job_type=sample(LETTERS[1:10],1000, replace=T,prob=prob),
                    gender=sample(c("male","female"), 1000, replace=T))

# tabulando

datax <- data1 %>% group_by(job_type,gender) %>% 
  summarise(frec=n()) %>% mutate(frec=if_else(gender=="male",-frec,frec))

tdf <- datax %>% filter(gender=="female") %>% arrange(frec)
order <- tdf$job_type

tdf <- tdf %>% mutate(prop=round(frec/404,digits = 2))

datax %>% ggplot(aes(x=job_type,y=frec,group=gender,fill=gender))+
  geom_bar(stat = "identity",width = 0.95)+
  coord_flip() + scale_x_discrete(limits=order) +
  scale_y_continuous(breaks = seq(-150,150,50),
                     labels = abs(seq(-150,150,50))) +
  geom_label(data = datax, aes(x=job_type,y=frec,
                 label=scales::percent(frec,scale = 0.05)),
             fill="gray90",alpha=0.7,size=3.25)

1.29 scale_x_dicrete() y scale_y_discrete()

F <- bio %>% ggplot(aes(x=especies,y=carbono, fill=especies))+
  geom_boxplot()

F + scale_x_discrete(labels=c("A. smithii"="AS",
                              "A. tamarugo"="AT",
                              "B. myrtifolia"="BM",
                              "C. fraseri"="CF",
                              "E. botryoides"="EB",
                              "E. maculata"="EMA",
                              "E. pilularis"="EP"))

# Forma de abreviado sencilla:

F + scale_x_discrete(labels=abbreviate(names.arg = unique(bio$especies),
                                       minlength = 4))

1.30 scale_fill_brewer() y scale_color_brewer()

N <- bio %>% 
  ggplot(aes(x=especies,y=carbono,fill=especies))+
  geom_boxplot()

G <- bio %>% 
  ggplot(aes(x=especies,y=carbono,color=especies))+
  geom_boxplot()

# Aplicando el rellenado y coloreado de figuras con colores brewer:
### RColorBrewer::brewer.pal.info

N + scale_fill_brewer(palette = "Dark2")

G + scale_color_brewer(palette = "Accent")

1.31 scale_fill_color_tableau() y scale_color_tableau()

Requiere la librería ggthemes()

N + scale_fill_tableau(palette="Summer")

G + scale_color_tableau(palette="Winter")

1.32 scale_color_grey() y scale_fill_grey()

G + scale_color_grey()

N + scale_fill_grey()

N + scale_fill_grey(start = 0.35, end = 0.8)

1.33 scale_fill_manual()

N + scale_fill_manual(values = c("#f0bc00","#b2b6f5","#1c3080",
                                          "#361757","#461757",
                                          "#690f42","#165e0c"))

1.34 scale_fill_gradientn()

antropo <- read.xlsx("anthropometry.xlsx")
H <- antropo %>% ggplot(aes(x=foot_length, y=height)) + 
  geom_hex()

# Creamos una paleta:
paleta <- c("#0D0887FF","#7E03A8FF","#CC4678FF",
                       "#F89441FF")
                       
H + scale_fill_gradientn(colours = viridisLite::cividis(10))

H + scale_fill_gradientn(colors = viridisLite::inferno(10))

H + scale_fill_gradientn(colors = viridisLite::magma(10))

H + scale_fill_gradientn(colors = viridisLite::plasma(10))

H + scale_fill_gradientn(colors = heat.colors(10))

H + scale_fill_gradientn(colors = paleta)

H + scale_fill_gradientn(colors = c("#f0bc00","#b2b6f5","#1c3080",
                                    "#361757","#461757"))

H + scale_fill_gradientn(colours=topo.colors(7),
                         breaks=c(1,2,600))

1.35 scale_fill_gradient()

H + scale_fill_gradient(low = "red",high = "yellow")

H + scale_fill_gradient(low = "blue",high = "lightgreen")

1.36 scale_fill_gradient2()

H + scale_fill_gradient2(low = "darkgreen", mid = "gray80",
                         high = "lightyellow",midpoint = 30)

1.37 Formas de puntos y lineas (argumento shape)

1.37.1 scale_shape_manual()

msleep <- msleep %>% filter(!is.na(vore))

M <- msleep %>% ggplot() + 
            geom_point(aes(x=sleep_total,y=sleep_rem,
                           shape=vore,color=vore),
                       size=3,alpha=0.7)

M + geom_smooth(aes(x=sleep_total,y=sleep_rem,
                    lty=vore,color=vore),lwd=1,method = "lm",
                se=FALSE)

M + scale_shape_manual(values = c(3,6,15,20)) + 
  scale_color_brewer(palette = "Dark2")

1.38 scale_type_manual()

linea <- read.xlsx("lines.xlsx")

Z <- linea %>% ggplot()+
  geom_line(aes(x=conc.enzima,y=actividad,
                color=z,lty=z),lwd=1.3) +
  scale_x_continuous(breaks = seq(0,80,10))


Z + scale_linetype_manual(values = c(3,5))

1.39 facet_wrap()

G <- iris %>% gather(key=variables,value=valor,1:4) %>% 
  ggplot(aes(x=variables,y=valor,fill=Species,
             color=Species)) + 
  geom_boxplot() + scale_fill_brewer(palette = "PuOr")

G + coord_flip() + scale_x_discrete(labels=abbreviate)

G + coord_flip() + scale_x_discrete(labels=abbreviate) +
  facet_wrap(~Species)

G + facet_wrap(~Species)

G + coord_flip() + scale_x_discrete(labels=abbreviate) +
  facet_wrap(~Species, nrow = 3)

G + coord_flip() + scale_x_discrete(labels=abbreviate) +
  facet_wrap(~Species, ncol = 2)

1.40 facet_grid()

library(dslabs)

gapminder$ferticateg <- cut(gapminder$fertility,
                            breaks = c(-Inf,2,4,
                                       6,8,Inf),
                            labels = c("<2",
                                       "2-4",
                                       "4-6",
                                       "6-8",
                                       ">8"))

gapminder <- na.omit(gapminder)

X <- gapminder %>% ggplot(aes(x=infant_mortality,
                              y=life_expectancy,
                              color=continent))+
  geom_point()


X + facet_grid(.~continent)

X + facet_grid(continent~.)

X + facet_grid(.~ferticateg)

X + facet_grid(ferticateg~.)

X + facet_grid(ferticateg~continent)

## Usando un poco de facet_wrap():

X + facet_wrap(~continent)

X + facet_wrap(~continent, ncol=5)

X + facet_wrap(ferticateg~continent,
               ncol = 3, nrow = 7)

1.41 labs()

J <- G + labs(title="Pétalos por especie",
              subtitle = "Valores de longitud",
              x="Características florales",
              y="Medición (cm)",
              caption = "Elaborado por:\nBryan Quispe")

# Cambiando el nombre de la leyenda:
J + labs(fill="Flores",color="Flores") + 
  scale_fill_brewer(palette="PuOr") +
  scale_color_brewer(palette="Dark2")

1.42 Usando ggThemeAssist()

J + theme(plot.subtitle = element_text(size = 12,
    face = "italic", hjust = 0.5), plot.caption = element_text(size = 14,
    face = "bold"), panel.grid.major = element_line(colour = "gray30"),
    panel.grid.minor = element_line(linetype = "blank"),
    axis.text = element_text(family = "serif",
        face = "italic"), plot.title = element_text(hjust = 0.5),
    legend.title = element_text(face = "bold",
        family = "Helvetica"), 
    panel.background = element_rect(fill = "palegoldenrod",
        colour = "gray70"), 
    plot.background = element_rect(fill = "cornsilk4",
        colour = "black"), legend.key = element_rect(fill = NA),
    legend.background = element_rect(fill = NA),
    legend.position = "top", legend.direction = "horizontal")

1.43 Gráficos de alta calidad

## png:
png("msleep.png", res = 600,units = "cm",width = 20,
    height = 14, bg="black")

M + geom_label_repel(aes(x=sleep_total,y=sleep_rem,
                         label=vore,fill=vore),alpha=0.25)

dev.off()

## jpeg:
jpeg("msleep.png", res = 600,units = "cm",width = 20,
    height = 14, bg="black")

M + geom_label_repel(aes(x=sleep_total,y=sleep_rem,
                         label=vore,fill=vore),alpha=0.25)

dev.off()

## pdf:

pdf("mypdf.pdf",width = 8,height = 4,bg="white",
    colormodel = "cmyk",paper = "A4")

M + geom_label_repel(aes(x=sleep_total,y=sleep_rem,
                         label=vore,fill=vore),alpha=0.25)

dev.off()

1.44 Adición de texto de correlaciones (R y p-value)

df <- read.xlsx("multiplicata_fvs_choice_updated2020.xlsx")

df2 <- df %>% filter(!yr_collected %in% 2007:2009)
df2 <- df %>% filter(yr_collected %in% 2010:2019)

A <- df2 %>% ggplot(aes(x=svl_hi,y=mass_hi))+
  geom_point(alpha=0.3,color="coral", size=3)+
  facet_wrap(~yr_collected,ncol = 2)+
  labs(x="Longitud Hocico Respiradero (SLV)",
       y="Condición corporal (mass/SLV",tag = "A") +
  ggpubr::stat_cor(aes(label=paste(..r.label..,
                                   ..p.label..,
                                   sep="~~")),color="black",
                   geom = "text", label.x = 35,label.y = 20,
                   size=3.5);A

1.45 Adicionar regresiones con texto:

B <- df2 %>% ggplot(aes(x=svl_hi,y=mass_hi,
                        color=factor(yr_collected)))+
  geom_point(alpha=0.3,size=3) + geom_smooth(method = "lm",
                                             se=FALSE) + 
  facet_wrap(~yr_collected,ncol = 2) + 
  labs(x="Longitud Hocico Respiradero (SLV)",
       y="Condición corporal (mass/SLV",tag = "B") +
  ggpubr::stat_regline_equation(label.x = 35,label.y = 20,
                                aes(label=paste(..eq.label..,
                                                ..adj.rr.label..,
                                                sep="~~~")));B

1.46 Patchwork: unión de gráficos

C <- iris %>% ggplot(aes(Petal.Length,Sepal.Length)) + 
  geom_point()

D <- iris %>% ggplot(aes(Petal.Length,Species)) +
  geom_boxplot()

# Usando la librería:
library(patchwork)
A+B

A+B+C+D

A+B|C+D

A|C|D

C+D|A

C/D

(C+D)/B

#Cambios especiales: operador &
(C/D|A)&
  theme(plot.background = 
          element_rect(fill = "gray90"))

(A|B)&theme(plot.background = 
              element_rect(fill = "gray90"))

ggpubr::ggarrange(A,B)