Remake of plot from https://www.kdnuggets.com/data-science-salaries-job-market-analysis-2024-2025

data-science-salaries.R

# Remake of plot from https://www.kdnuggets.com/data-science-salaries-job-market-analysis-2024-2025

# load packages

library(tidyverse)
library(scales)

# create data frame

ds_salaries <- tribble(
  ~experience, ~salary, ~label,
  0.5,  117276, "0-1 years",
    2,  128403, "1-3 years",
    5,  141390, "4-6 years",
    8,  152966, "7-9 years",
    12, 166818, "10-14 years",
    15, 189884, "15+ years"
)

# plot

ggplot(
  ds_salaries,
  aes(x = experience*3, y = salary)
) +
  geom_col(width = 4, fill = "#e8c95c") +
  labs(
    x = "Median yearly total compensation for data scientists in the US by years of experience (Glassdoor)",
    y = NULL,
    title = "Data Science Salaries\nby Experience Level"
  ) +
  geom_text(
    aes(y = salary + 20000, label = str_to_upper(label)), 
    color = "lightgray", fontface = "bold", size = 3
  ) +
  geom_text(
    aes(y = salary + 10000, label = dollar(salary)), 
    color = "#75b6be", fontface = "bold", size = 4
  ) +
  theme_void() +
  theme(
    plot.background = element_rect(fill = "black"),
    plot.title = element_text(color = "#e8c95c", face = "bold", hjust = 0.5, size = 25),
    axis.title.x = element_text(color = "white", size = 10, margin = margin(t = 2.5, b = 2.5, l = 0, r = 0))
  )

The original plot from KD Nuggets:

The updated plot with code above, scaling x and y axes correctly: