plot_benchmark_ranger.R

# License: BSD (3-clause)
# Author: Denis A. Engemann <[email protected]>

library(ggplot2)
library(tidymodels)
library(readr)
library(wesanderson)

hotels <-
  read_csv('https://tidymodels.org/start/case-study/hotels.csv') %>%
  mutate_if(is.character, as.factor)

dat <- list.files(pattern = "*.csv") %>%
  map_df(read_csv)

dat$n_cores <- dat$expr %>% 
  substr(7, 7) %>% 
  as.numeric()

dat$machine <- case_when(
  dat$os == "macOS Catalina 10.15.6" ~ "MBP-15 2017 i7",
  dat$os == "macOS  11.1" ~ "MBA 2020 M1",
  dat$os == "macOS Big Sur 10.16" ~ "MBA 2020 M1",
)

dat$label <- paste(dat$machine, dat$arch)

n_samp <- nrow(hotels) * 0.8
title <-
  sprintf("Hotels, n=%d, RF (ranger) mtry=8, min_n=7, trees=100",
          n_samp)

dat %>%
  mutate_if(is.character, as.factor) %>%
  ggplot(aes(x = n_cores, y = time / 1e9, color = label)) +
  geom_jitter() +
  geom_smooth(show.legend = F) +
  theme_classic() +
  theme(legend.position = c(0.8, 0.8),
        text = element_text(size = 14, family = "Helvetica")) +
  scale_y_continuous(breaks = 1:12) +
  labs(x = "Number of CPUs", y = "Time [sec.]",
       title = title) +
  scale_color_manual(
    values = wes_palette(n = 5, name = "Zissou1")[c(1, 3, 5)])

ggsave("ranger_bench.png", dpi = 300)