Fit a different model for each row of a list-columns data frame

Question

What is the best way to fit different model formulae that vary by the row of a data frame with the list-columns data structure in tidyverse?

In R for Data Science, H

Answer 1

I found purrr::modify_depth() that does what I want to do with est_model() in the original question. This is the solution I am now happy with:

library(gapminder)
library(tidyverse)
library(purrr)
library(broom)

fmlas <- tibble::tribble(
  ~continent, ~formula,
  "Asia", ~lm(lifeExp ~ year, data = .),
  "Europe", ~lm(lifeExp ~ year + pop, data = .),
  "Africa", ~lm(lifeExp ~ year + gdpPercap, data = .),
  "Americas", ~lm(lifeExp ~ year - 1, data = .),
  "Oceania", ~lm(lifeExp ~ year + pop + gdpPercap, data = .)
)

by_continent <- gapminder %>% 
  nest(-continent) %>%
  left_join(fmlas) %>%
  mutate(model=map2(data, formula, ~modify_depth(.x, 0, .y)))

by_continent %>% 
  mutate(glance=map(model, glance)) %>% 
  unnest(glance, .drop=T)