Issue applying purrr::map() inside nested dataframe

Question

I'm working through chapter 25 "Many Models" of Hadley Wickham's R for Data Science https://r4ds.had.co.nz/many-models.html , however I'm running into issues in recreating the examples in 25.2.2.

Here's what I have so far (and what's working):

require(gapminder); require(tidyverse); require(broom); require(modelr)

by_country <- gapminder %>% group_by(country,continent) %>% nest()
head(by_country)

# A tibble: 6 x 3
  country     continent data             
  <fct>       <fct>     <list>           
1 Afghanistan Asia      <tibble [12 × 4]>
2 Albania     Europe    <tibble [12 × 4]>
3 Algeria     Africa    <tibble [12 × 4]>
4 Angola      Africa    <tibble [12 × 4]>
5 Argentina   Americas  <tibble [12 × 4]>
6 Australia   Oceania   <tibble [12 × 4]>

Then defining the lm() to apply to each of country's set of data:

country_model <- function(df) {
  lm(lifeExp ~ year, data = df)
}

And then this next line doesn't work:

by_country <- by_country %>%
  mutate(model = map(data,country_model))

with error message

Error in eval(predvars, data, env) : object 'lifeExp' not found

Despite, to my eyes, what I've written being the same as what's appeared in Hadley's chapter.

I'm unsure if this is a recent issue that used to work, as someone else has apparently had an issue with the example: https://github.com/hadley/r4ds/issues/766 (with no solution)

Any help would be greatly appreciated!

I've got dplyr v0.8.3 and purrr v0.3.2 and this runs just fine: I'm getting a data frame with a list-column of nested data and a list-column of lm objects. I think I'm a little behind on my updates though — camille
– camille, Commented Sep 22, 2019 at 15:25
@Parfait good idea, though it results in the following error: Error in UseMethod("unnest_") : no applicable method for 'unnest_' applied to an object of class "list" — yanqui
– yanqui, Commented Sep 22, 2019 at 16:46
@Aurèle my sessioninfo is linked here: pastebin.com/t0WhrURN — yanqui
– yanqui, Commented Sep 22, 2019 at 16:49
I am able to run the example. The only difference is that I call libraries instead of require, i.e. library(gapminder) library(tidyverse) library(broom) library(modelr) — Vitali Avagyan
– Vitali Avagyan, Commented Sep 22, 2019 at 17:56

W148SMH · Accepted Answer · 2019-09-24 00:45:59Z

1

You don't need to redefine "by_country" twice.

country_model <- function(df) {
lm(lifeExp ~ year, data = df)
}

by_country <- gapminder %>% 
group_by(country,continent) %>% 
nest()%>%
mutate(model = map(data,country_model))

answered Sep 24, 2019 at 0:45

W148SMH

1721 silver badge12 bronze badges

Sign up to request clarification or add additional context in comments.

Comments

Ethan Nguyen · Accepted Answer · 2020-09-08 08:55:47Z

1

The group_by + nest combo can be replaced with nest_by, only thing different is the results is grouped rowwise so you would need to put functions in a list. Something like this:

results <- gapminder %>% 
    nest_by(continent, country) %>% 
    mutate(model = list(lm(lifeExp ~ year, data = data)))

answered Sep 8, 2020 at 8:55

Ethan Nguyen

565 bronze badges

Collectives™ on Stack Overflow

Issue applying purrr::map() inside nested dataframe

2 Answers 2

Comments

Comments

Your Answer

Linked

Hot Network Questions

Collectives™ on Stack Overflow

2 Answers 2

Comments

Comments

Your Answer

Sign up or log in

Post as a guest

Linked

Related