13  Aufgabe Woche 6

# Übungen Woche 6

pacman::p_load("sjPlot", "texreg", "haven", "readr", "readxl", "fixest", "MASS",
               "glm.predict", "scales", "tidyverse")

nsim = readRDS("nsim.rds")

plot_a = nsim %>% 
  ggplot(aes(x = Teilnahme, y = Vertrauen, color = Gender)) +
  geom_point() +
  scale_color_manual(values = c("Mann" = "blue4", "Frau" = "red4"), # Benutzerdefinierte Farben, siehe bspw. https://sape.inf.usi.ch/quick-reference/ggplot2/colour
                     labels = c("weiblich", "männlich")) + #geom_point = Streudiagramm
  theme_minimal() +
  labs(x = "Teilnahme", y = "Vertrauen", color = "Geschlecht")

lm_nsim = lm(Teilnahme ~ Vertrauen, nsim)
lm_nsim

plot_a

# Variablen hängen negativ zusammen: Je mehr Teilnahme, desto mehr Vertrauen

plot_b <- plot_a +
  geom_abline(slope = lm_nsim$coefficients[2], intercept = lm_nsim$coefficients[1],
              color = "red")
plot_b

nsim %>% 
  group_by(Gender) %>% 
  summarize(Mean = mean(Teilnahme, na.rm = TRUE))

nsim %>% 
  group_by(Gender) %>% 
  summarize(Mean = mean(Vertrauen, na.rm = TRUE))

poisson = read_xlsx(choose.files())

poisson = poisson %>% 
  mutate(noyades = noyades / population)

poisson = poisson %>% 
  mutate(poissoneries = poissonneries / population)

lm_poi = lm(noyades ~ poissonneries, poisson)
lm_poi

lm_poi_mer = lm(noyades ~ poissonneries + mer, poisson)
lm_poi_mer

screenreg(list(lm_poi, lm_poi_mer))

poisson1 = poisson %>% 
  mutate(mer = ifelse(mer == 1, "Zugang", "Kein Zugang"))

plot_poisson = poisson1 %>% 
  ggplot(aes(x = poissonneries, y = noyades, color = mer)) +
  geom_point() +
  theme_bw() +
  labs(x = "poissonneries", y = "noyades", color = "Zugang zu Meer")

plot_poisson

plot_poisson2 = plot_poisson + geom_abline(slope = lm_poi$coefficients[2], intercept = lm_poi$coefficients[1],
                                           color = "red")
plot_poisson3 = plot_poisson2 +   geom_abline(slope = lm_poi_mer$coefficients[2], intercept = lm_poi_mer$coefficients[1],
                                              color = "green")

plotpoisson4 = plot_poisson3 + geom_abline(intercept = coef(lm_poi)[1] + coef(lm_poi_mer)[3], slope = coef(lm_poi)[2], color = "orange")

lm_poi
lm_poi_mer

plotpoisson4

plot_poisson3

screenreg(list(lm_poi, lm_poi_mer))

plot_poisson3

lm5 = lm(noyades ~ poissonneries, poisson)
lm6 = lm(noyades ~ poissonneries + mer, poisson)
csim = read_dta("csim.dta")


csim

lm_csim = lm(y ~ x1, csim)

## Aufgabe im Tutorat Geschaft mit Chatgpt -> Quarto file erweitern.

lm_csim2 = lm(y ~ x1 + x2, csim)


coef(lm_csim)["x1"]
coef(lm_csim2)["x1"]
csim2 = csim %>% 
  mutate(x4 = 0.15 * csim$y + rnorm(nrow(csim), mean = 3, sd = 6))


lm_csim3 = lm(y ~ x1 + x4, csim2)                

coef(lm_csim3)["x1"]
coef(lm_csim2)["x1"]