Plotcon 2016 Hadley Wickham

All script is copied from Hadley Wickham’s Plotcon 2016 lecture. Video Link

Load library

library(tidyverse)
library(gapminder)

Get Data

head(gapminder)

country	continent	year	lifeExp	pop	gdpPercap
Afghanistan	Asia	1952	28.801	8425333	779.4453
Afghanistan	Asia	1957	30.332	9240934	820.8530
Afghanistan	Asia	1962	31.997	10267083	853.1007
Afghanistan	Asia	1967	34.020	11537966	836.1971
Afghanistan	Asia	1972	36.088	13079460	739.9811
Afghanistan	Asia	1977	38.438	14880372	786.1134

gapminder <- gapminder %>% mutate(year1950 = year - 1950)

Nested data

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

# head(by_country)
by_country$data[[1]] %>% head() # Afghanistan

year	lifeExp	pop	gdpPercap	year1950
1952	28.801	8425333	779.4453	2
1957	30.332	9240934	820.8530	7
1962	31.997	10267083	853.1007	12
1967	34.020	11537966	836.1971	17
1972	36.088	13079460	739.9811	22
1977	38.438	14880372	786.1134	27

by_country$data[[2]] %>% head() # Albania

year	lifeExp	pop	gdpPercap	year1950
1952	55.23	1282697	1601.056	2
1957	59.28	1476505	1942.284	7
1962	64.82	1728137	2312.889	12
1967	66.22	1984060	2760.197	17
1972	67.69	2263554	3313.422	22
1977	68.93	2509048	3533.004	27

Fit models

models <- by_country %>% 
  mutate(model = data %>% map(function(x) lm(lifeExp ~ year1950, data = x)))


# models %>% filter(continent == "Africa") %>% head()

Broom

models <- models %>%
  mutate(
    glance = model %>% map(broom::glance),
    rsq = glance %>% map_dbl("r.squared"),
    tidy = model %>% map(broom::tidy),
    augment = model %>% map(broom::augment)
    
  )

# models %>% arrange(desc(rsq)) %>% head()


models[sample(1:142, 15),] %>% ggplot(aes(rsq, reorder(country, rsq))) +
  geom_point(aes(colour = continent))

Unnest

# 
# models %>% unnest(data)
# models %>% unnest(glance, .drop = TRUE)
# models %>% unnest(tidy)

models %>%
  unnest(tidy) %>%
  select(continent, country, term, estimate, rsq) %>%
  spread(term, estimate) %>% 
  ggplot(aes(`(Intercept)`, year1950)) +
  geom_point(aes(colour = continent, size = rsq)) +
  geom_smooth(se = FALSE) +
  xlab("Life Expectancy (1950)") +
  ylab("Yearly improvement") +
  scale_size_area()

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