Labels for rivers in the ggplot2 framework

ggriverlab aims to introduce labels that follow lines to creat nice maps with ggplot2. While this function is common in GIS software, I could not find satisfying solutions for R plotting, so here is my take on it.

Installation

The package can be installed from github with the help of the devtool package.

# install.packages("devtools")
devtools::install_github("Clement-Viguier/ggriverlab")

Demo

library(ggriverlab)
library(riverdist)
library(ggplot2)
library(dplyr)


river <- as.data.frame(Gulk$lines[4])
colnames(river) <- c("x", "y")

ggplot(river, aes(x,y)) + geom_path() + coord_fixed()
ggplot(river, aes(x,y)) + geom_path() + coord_fixed() + geom_river_label(aes(label = "Gulk", offset = 0.24), reverse = F)

?geom_river_label

# Plotting the full river

branches <- 1:14
branch_list <- lapply(branches, function(b, Gulk){
  river <- as.data.frame(Gulk$lines[b])
  colnames(river) <- c("x", "y")
  river$branch <- b
  return(river)
}, Gulk)
full_river<- do.call(rbind, branch_list)

ggplot(full_river, aes(x,y, group = branch, colour = as.factor(branch))) + geom_path() + coord_fixed() + geom_river_label(aes(label = "Gulk"), check_length = F) +
 theme_minimal() + theme(legend.position = "none")

# Check_length avoids to print labels that cannot fit the river portion.
ggplot(full_river, aes(x,y, group = branch, colour = as.factor(branch))) + geom_path() + coord_fixed() + geom_river_label(aes(label = "Gulk"), check_length = T) +
  theme_minimal() +theme(legend.position = "none")

ggplot(full_river, aes(x,y, group = branch, colour = as.factor(branch))) + geom_path() + coord_fixed() + geom_river_label(aes(label = "Gulk"), offset = 0.15, dist = 0.01, check_length = T) +
  theme_minimal() +theme(legend.position = "none")

The geom_river_label function can be used to plot multiple river, or branches at once. If they are too short, the label can be exclude thanks to the option check_length = T. The above code produce the following map of the river:

# install.packages('devtools')
# devtools::install_github('thomasp85/gganimate')
library(gganimate)

# generate data for the animation
offsets <- seq(0.05, 0.9, length.out = 20)
list_data <- lapply(offsets, function(os, river){
  river$off <- os
  return(river)
}, river)
data_offset <- do.call(rbind, list_data)

anim <- ggplot(data_offset, aes(x,y)) + geom_path() + coord_fixed() + geom_river_label(aes(label = "Gulk", offset = off), relative = T, check_length = F, reverse = F, repeated = F)+
  transition_states(off, 1, 3)
anim
anim_save("Images/moving_label_smooth.gif", last_animation())

The effect of the different parameters on the label position and shape can be seen here:

Larger scale data

Oceans and rivers

library(raster)
library(tidyr)

ne_rivers <- shapefile("./data/ne_50m_rivers_lake_centerlines.shp")
ne_lands  <- shapefile("./data/ne_50m_land.shp")
ne_lakes  <- shapefile("./data/ne_50m_lakes.shp")

data <- ne_rivers@data
data$id <- rownames(data)
rivers_df <- fortify(ne_rivers) %>% left_join(data, by = "id")

lands <- fortify(ne_lands)


ggplot(rivers_df, aes(long, lat, group = group)) +
   geom_path(aes(), colour = "steelblue") + geom_river_label(aes(label = label), colour = "steelblue") +
  coord_fixed() + 
  facet_wrap(~id, scales = "free")+
  theme_linedraw()+
  theme_classic() + theme(panel.background = element_rect(fill = "steelblue"))

TO DO LIST

• ✓ make it work with point: place points along the path in place of letters;
• ✓ include in geom;
• □ make it work with entire labels as an option;
• ✓ make it work with letters;

TO DO if the rest is done and clean

• □ make river specific geom (size~order);
• □ make interactive tool to test the parameters;
• ✓ find good default parameters;
• □ scale against text size or plot size ?;