broom_odds_ratios.md

library(tidyverse)
library(broom)

# Basic logit predicting if a car is an SUV or not
cars_binary <- mpg %>% 
  # Make new column marking if it's an SUV or not
  mutate(suv = class == "suv")

model <- glm(suv ~ cty + displ, 
             data = cars_binary,
             family = binomial(link = "logit"))

# These are log odds
tidy(model, conf.int = TRUE)
#> # A tibble: 3 x 7
#>   term        estimate std.error statistic   p.value conf.low conf.high
#>   <chr>          <dbl>     <dbl>     <dbl>     <dbl>    <dbl>     <dbl>
#> 1 (Intercept)    4.42     2.12       2.09  0.0367       0.371     8.74 
#> 2 cty           -0.384    0.0931    -4.13  0.0000369   -0.577    -0.209
#> 3 displ          0.132    0.219      0.603 0.546       -0.306     0.561

# These are exponentiated odds ratios
tidy(model, exponentiate = TRUE, conf.int = TRUE)
#> # A tibble: 3 x 7
#>   term        estimate std.error statistic   p.value conf.low conf.high
#>   <chr>          <dbl>     <dbl>     <dbl>     <dbl>    <dbl>     <dbl>
#> 1 (Intercept)   83.4      2.12       2.09  0.0367       1.45   6235.   
#> 2 cty            0.681    0.0931    -4.13  0.0000369    0.562     0.811
#> 3 displ          1.14     0.219      0.603 0.546        0.737     1.75

# Coefficient plot of log odds
coefs_to_plot_log_odds <- tidy(model, conf.int = TRUE) %>% 
  filter(term != "(Intercept)")

ggplot(coefs_to_plot_log_odds, aes(x = estimate, y = term)) +
  geom_vline(xintercept = 0, color = "red") +
  geom_pointrange(aes(xmin = conf.low, xmax = conf.high)) +
  labs(x = "Coefficient (log odds)")

# Coefficient plot of odds ratios
coefs_to_plot_or <- tidy(model, exponentiate = TRUE, conf.int = TRUE) %>% 
  filter(term != "(Intercept)")

ggplot(coefs_to_plot_or, aes(x = estimate, y = term)) +
  geom_vline(xintercept = 1, color = "red") +
  geom_pointrange(aes(xmin = conf.low, xmax = conf.high)) +
  labs(x = "Coefficient (odds ratio)")

^{Created on 2020-05-19 by the reprex package (v0.3.0)}