Tag 6 Folgt

6.1 Lösung Tag 5

library(tidyverse)
library(broom)

6.1.1 Mit den Datasaurus Dozen Datensets

datasauRus::datasaurus_dozen

  • Visualisiere alle Sets gemeinsam in einem ggplot Scatterplot, nutze dazu facet_wrap.

  • Füge mittels geom_smooth lineare Trendlinien hinzu

  • Fitte eine Lineare Regression and jedes der Datensets. Nutze dazu die Techniken aus R4DataScience: Many Models und das broom package.

  • Analysiere die Fits.

dinos <- datasauRus::datasaurus_dozen
dinos
dinos %>% 
  count(dataset)
dinos %>% 
  ggplot(aes(x = x, y = y)) +
  geom_point() +
  geom_smooth(method = "lm") +
  facet_wrap("dataset", scales = "free_y") +
  ggthemes::theme_few()
nested_dinos <- dinos %>% 
  group_nest(dataset)
nested_dinos
beispiel <- nested_dinos$data[[1]]

model <- lm(y ~ x, data = beispiel)
summary(model)

glance(model)
tidy(model)
coef(model)
all_models <- nested_dinos %>% 
  mutate(
    model        = map(data, ~ lm(y ~ x, data = .x)),
    glance       = map(model, glance),
    coefficients = map(model, coef),
    corr         = map(data, ~ cor(.x$x, .x$y))
  )
all_models
all_models %>% 
  select(dataset, coefficients) %>% 
  unnest_wider(coefficients)
all_models %>% 
  select(dataset, corr) %>% 
  unnest_wider(corr)

6.2 Datenanalyse, Beispiel: Global Plastic Waste

6.3 Feedbackrunde