CODEWITHSUNDEEP
rplotggplot2
Abby_studies_fish
Abby_studies_fish

Reputation: 23

Trying to plot multiple plots with two factors

EDIT: SOLVED Used this code:

gg5 <- ggplot(df_tidy, aes(x=Year, y=Landing))+
  geom_line()
gg5 + facet_wrap(~df_tidy$country)

I'm sorry, this is slightly a duplicate question. But I'm asking it because I've tried the different answers and nothing seems to work.

I want to create some plots. I want to have separate plots per country. And I want to have different lines on each plot for region. I've got ggplot2 and I've got ggplus. In total, I'm going to have approximately 150 plots, so facet_multiple seems to be the answer so I can spread it over a number of pages. But what I am getting in practice is not what I anticipate in theory.

Here's some dummy data:

library(tidyverse)
country = c("USA", "Canada", "UK", "France", "Mexico")
region = c("North America", "North America", "Europe", "Europe", "North America")  
2000 = c(50, 100, 75, 60, 25) 
2001 = c(200, 50, 150, 125, 60)
2002 = c(125,75,60, 75, 25)
df = data.frame(country, region, 2000, 2001, 2002) 
df_tidy <- df %>% gather(Year, Landing, c(3:5))

     country     region      Year   Value
 1      USA   North America  2000    50
 2   Canada   North America  2000   100
 3       UK       Europe     2000    75
 4   France       Europe     2000    60
 5   Mexico   North America  2000    25
 6      USA   North America  2001   200
 7   Canada   North America  2001    50
 8       UK       Europe     2001   150
 9   France       Europe     2001   125
10   Mexico   North America  2001    60
11      USA   North America  2002   125
12   Canada   North America  2002    75
13       UK      Europe      2002    60
14   France      Europe      2002    75
15   Mexico   North America  2002    25

Please just imagine there might be 2 (or 3 or 4..) different regions per country.

Now, if I break individual countries out into their own data.frame, I get the exact plot I want: Correct plot!

#This works
USA <- df_tidy %>%
filter(Country %in% c("USA") & between(Year,2000,2002)) %>%
select(Country, Year, Landing)

ggplot(USA, aes(x=Year, y=Landing)) +
geom_line(data = USA)
#End of USA plot

But when I try to ggplot the entire data frame using facets, I get something like this: Really weird plot

I've tried multiple different ways to do it, and if I don't get an error message, I get the really weird plot.

Some code I've tried:

#Attempt 1:
gg1 <- ggplot(df_tidy, aes(x=df_tidy$Year, y=df_tidy$Landing, color = df_tidy$region))+
geom_line()

#doesnt work
gg1+facet_wrap(~df_tidy$country, ncol = 4, nrow = 4)

#doesnt work
gg1 + facet_grid(df_tidy$country) + facet_grid(df_tidy$region) 

#doesnt work   
facet_grid(gg1, facets = "country") 

#Attempt 2
 plist = lapply(split(df_tidy, df_tidy$country), function(d) {
 ggplot(d, aes(x=Year, y=Landing))+
 geom_line()+
 facet_wrap(~ df_tidy$country)+
 scale_y_continuous()})

library(gridExtra)
for(i in seq(1, length(plist), 6)) {
pdf(paste0("country_", i, ".pdf"), 7, 5)
grid.arrange(grobs=plist[i:(i+5)],
            ncol(3, left="Landing", bottom="Year"))}

Anyways, I can put up every attempt, but I've deleted and changed code so much I don't even remember where I started. Any advice would be appreciated because I really dont want to have to each country by seperate data frame. Thanks!

Update - after trying the solution, here's what I'm still getting:

myPlots <- list()
 for(i in unique(df_tidy$region)) {
   myPlots[[i]] <- ggplot(subset(df_tidy, df_tidy$country == i),
                      aes(Year, Landing, color = region)) +
 geom_line(size=2)+
 labs(title = paste("Landings in", i),
      subtitle = "From 2000-2002",
      x = "Year",
      y = "Landing",
      color = NULL) +
 scale_color_brewer(palette = "Set2") + 
 theme_minimal()+
 theme(legend.position = "bottom")
 }

library(ggpubr) ggarrange(plotlist = myPlots)

loop plot

ggplot(df_tidy, aes(Year, Landing))+
  geom_line()+
  facet_wrap(~ df_tidy$region + df_tidy$country)

facet_wrap

Upvotes: 1

Views: 987

Answers (1)

pogibas
pogibas

Reputation: 28339

As you mentioned this question might be a duplicate as there have been many similar problems, however sometimes they are very specific.

To begin with we can use simple facet_wrap:

ggplot(df_tidy, aes(Year, Landing)) +
    geom_line() +
    facet_wrap(~ region + country)

Here I'm creating facets that are combinations of region and country (~ region + country)

enter image description here

Another solution would be to

  1. Loop over regions and create separate plots
  2. Combine plots with some function

Here I will use same pallet per region (also added minimal visual tweaks).

# Loop over regions and create separate plots
myPlots <- list()
for(i in unique(df_tidy$region)) {
    myPlots[[i]] <- ggplot(subset(df_tidy, region == i), 
        aes(Year, Landing, color = country)) +
    geom_line(size = 2) +
    labs(title = paste("Landing change in", i),
         subtitle = "From xxxx to yyyy",
         x = "Year",
         y = "Landing",
         color = NULL) +
    scale_color_brewer(palette = "Set2") +
    theme_minimal() +
    theme(legend.position = "bottom")
}

# Combine plots with some function
library(ggpubr)
ggarrange(plotlist = myPlots)

enter image description here

Data & Packages:

library(tidyverse)
df_tidy <- structure(list(country = structure(c(5L, 1L, 4L, 2L, 3L, 5L, 
1L, 4L, 2L, 3L, 5L, 1L, 4L, 2L, 3L), .Label = c("Canada", "France", 
"Mexico", "UK", "USA"), class = "factor"), region = structure(c(2L, 
2L, 1L, 1L, 2L, 2L, 2L, 1L, 1L, 2L, 2L, 2L, 1L, 1L, 2L), .Label = c("Europe", 
"North America"), class = "factor"), Year = c(2000, 2000, 2000, 
2000, 2000, 2001, 2001, 2001, 2001, 2001, 2002, 2002, 2002, 2002, 
2002), Landing = c(50, 100, 75, 60, 25, 200, 50, 150, 125, 60, 
125, 75, 60, 75, 25)), row.names = c(NA, -15L), .Names = c("country", 
"region", "Year", "Landing"), class = "data.frame")

Upvotes: 0

Related Questions