第 85 章 探索性数据分析-哺乳动物脑量与体重的关系

我在网上看到这张图

找到数据来源是一篇文章，我下载了数据，希望能重复这张图。

85.1 读取数据

library(tidyverse)

data <- readr::read_rds("./demo_data/brain_size.rds") %>% 
   janitor::clean_names()
data

## # A tibble: 70 × 3
##    species         body_weight_kg brain_weight_g
##    <chr>                    <dbl>          <dbl>
##  1 Arctic fox                3.38           44.5
##  2 Owl monkey                0.48           15.5
##  3 Mountain beaver           1.35            8.1
##  4 Cow                     465             423  
##  5 Grey wolf                36.3            19.5
##  6 Goat                     27.7           115  
##  7 Roe deer                 14.8            98.2
##  8 Guinea pig                1.04            5.5
##  9 Vervet                    4.19           58  
## 10 Chinchilla                0.43            6.4
## # ℹ 60 more rows

85.2 数据变换

tb <- data %>%
  filter(body_weight_kg > 0, brain_weight_g > 0) %>%
  mutate(
    across(where(is.numeric), log10)
  )
tb

## # A tibble: 69 × 3
##    species         body_weight_kg brain_weight_g
##    <chr>                    <dbl>          <dbl>
##  1 Arctic fox              0.529           1.65 
##  2 Owl monkey             -0.319           1.19 
##  3 Mountain beaver         0.130           0.908
##  4 Cow                     2.67            2.63 
##  5 Grey wolf               1.56            1.29 
##  6 Goat                    1.44            2.06 
##  7 Roe deer                1.17            1.99 
##  8 Guinea pig              0.0170          0.740
##  9 Vervet                  0.622           1.76 
## 10 Chinchilla             -0.367           0.806
## # ℹ 59 more rows

85.3 线性模型

m <- lm(brain_weight_g ~ body_weight_kg, data = tb)

df <- tb %>% modelr::add_residuals(m)

85.4 可视化

library(colorspace)

## Warning: package 'colorspace' was built under R version 4.2.2

df %>% 
  ggplot(aes(x = body_weight_kg, y = brain_weight_g)) +
  geom_smooth(method = "lm", se = TRUE, color = "gray60", fill = "gray70", linetype = "dashed") +
  geom_point(aes(fill = resid), color = "black", size = 4, shape = 21) +
  ggrepel::geom_text_repel(aes(label = species)) +
  #scale_x_log10() +
  #scale_y_log10() +
  ggthemes::theme_economist() +
  theme(
    legend.position = "none"
  ) +
  scale_fill_continuous_diverging(palette = "Green-Orange")

## `geom_smooth()` using formula = 'y ~ x'

## Warning: ggrepel: 29 unlabeled data points (too many overlaps). Consider
## increasing max.overlaps