categorize date using jenks natural breaks

jenks_natural_breaks.R

library(stringr)
library(stringi)
library(ggplot2)
library(viridis)
library(data.table)
library(BAMMtools) # fast calculation of jenks natural breaks
library(ggt)

# load data
  data("iris")

# create Natural Jenks function
jenks_natural <- function(data, var, breaks){
  
  # data <- copy(iris)
  # var <- "Petal.Length"
  # breaks <- 5
 

# conver df to data.table  
  setDT(data)

# name of new column
  newvar <- paste0(var,"_jenks")

# calculate jenks natural breaks
  data[, paste0(newvar) := as.character(cut(get(var), breaks= getJenksBreaks(get(var), breaks), include.lowest = TRUE, dig.lab=3)) ]

# Edit factor text
  data[, paste0(newvar) := str_replace_all(get(newvar), "\\[|\\(|\\]", "") ]
  data[, paste0(newvar) := stri_replace_all_regex(get(newvar), "[,]", " - ") ]

# get factor labels
jenks_labels  <- data[, get(newvar)]  %>% table %>% names() %>% sort(decreasing = F) 

# recode variable
  data[, paste0(newvar) := factor(get(newvar), levels = jenks_labels)]
  
  return(data)
}




# apply function
  iris <- jenks_natural(iris, "Petal.Length", 5)





  
# plot using continuous values
  ggplot(iris, aes(x=Sepal.Length, y=Petal.Length, color=Petal.Length)) +
    scale_color_viridis( ) +
    geom_point() +
    theme_bw()

# plot using discrete values
  ggplot(iris, aes(x=Sepal.Length, y=Petal.Length, color=Petal.Length_jenks)) +
    scale_color_viridis( discrete = T) +
    geom_point() +
    theme_bw()