gapminder-eda.R

library(tidyverse)
library(gapminder)

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

# Nested data -------------------------------------------------------------

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

by_country
str(by_country)
by_country$data[[1]]
by_country[1, ]
by_country$data[[2]]

# Fit models --------------------------------------------------------------

country_model <- function(df) {
  lm(lifeExp ~ year1950, data = df)
}

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

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

# 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

models %>% arrange(desc(rsq))
models %>% filter(continent == "Africa")

models %>%
  ggplot(aes(rsq, reorder(country, rsq))) +
    geom_point(aes(colour = continent))

source("gapminder-shiny.R")

# Unnest ------------------------------------------------------------------

models
unnest(models, data) # back to where we started
unnest(models, glance, .drop = TRUE) %>% View()
unnest(models, 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()

models %>%
  unnest(augment) %>%
  ggplot(aes(year1950, .resid)) +
    geom_line(aes(group = country), alpha = 1/3) +
    geom_hline(yintercept = 0, colour = 'white', size = 2) +
    geom_smooth(se = FALSE) +
    facet_wrap(~continent)


# Initial summary plot ----------------------------------------------------

summary <- models %>%
  transmute(
    continent,
    country,
    slope = model %>% map(coef) %>% map_dbl(2),
    rsq = glance %>% map_dbl("r.squared")
  )

summary %>%
  ggplot(aes(rsq, slope)) +
  geom_point(aes(colour = continent)) +
  xlab(quote(R ^ 2)) +
  ylab("Estimated yearly increase in life expectancy") +
  theme(legend.position = "top", aspect.ratio = 1)

gapminder-shiny.R

library(shiny)
library(miniUI)

ui <- miniPage(
  miniButtonBlock(
    shiny::flowLayout(
      sliderInput("rsquared", "R^2", 0, 1, c(0, 0.25))
    )
  ),
  miniContentPanel(
    plotOutput("plot", height = "100%")
  )
)

server <- function(input, output) {
  selected_models <- reactive({
    models %>%
      filter(between(rsq, input$rsquared[1], input$rsquared[2])) %>%
      head(20)
  })

  rows <- reactive({
    ceiling(nrow(selected_models()) / 2)
  })

  output$plot <- renderPlot({
    selected_models() %>%
      semi_join(gapminder, ., by = "country") %>%
      ggplot(aes(year, lifeExp)) +
      geom_line() +
      facet_wrap(~country, ncol = 2) +
      theme(plot.margin = margin(0, 0, 0, 0))  +
      xlab(NULL) +
      ylab(NULL)
  }, height = function(...) rows() * 150, res = 96)
}

runGadget(ui, server, viewer = paneViewer("maximize"))

Kia Ora,
This is wonderful. Thank you for your time and work.
When I run source("gapminder-shiny.R") it throws back Warning: Error in UseMethod: no applicable method for 'margin' applied to an object of class "c('double', 'numeric')".
But if I remove or # this line: theme(plot.margin = margin(0, 0, 0, 0)) + from the ggplot() in the gapminder-shiny.R everything works properly.

BR,
Alex

Just found the youtube video on this, thanks alot really useful!!

in gapminder-eda.R, line 22, lm(lifeExp ~ year1950, data = df)
should this not be: lm(lifeExp ~ year1950, data = gapminder)
?