Código para la nota de medium sobre modelamiento en Ingeniería de Datos

modeladoinc.R

library(tidyverse)
library(haven)
theme_set(theme_bw(base_size = 24))


dat <- read_dta("http://fmwww.bc.edu/ec-p/data/wooldridge/wage1.dta")


dat %>% 
  ggplot() +
  aes(educ, wage) +
  geom_jitter() +
  scale_x_continuous(expand = c(0, 0)) +
  labs(x="Educación en años", 
       y= "Ingresos")


# Promedios locales
dat2 <- dat %>% 
  left_join(
    dat %>% 
      group_by(educ) %>% 
      summarise(avgwage = mean(wage)),
    by = "educ"
  )

dat2 %>% 
  ggplot() +
  aes(x= educ, y = wage) +
  geom_jitter() +
  geom_line(aes(y=avgwage), color = "blue", size = 1.2) +
  scale_x_continuous(expand = c(0, 0)) +
  labs(x="Educación en años", 
       y= "Ingresos")


# Lineal
lm(dat$wage~dat$educ))

dat2 %>% 
  ggplot() +
  aes(x= educ, y = wage) +
  geom_jitter() +
  scale_x_continuous(expand = c(0, 0)) +
  geom_smooth(method = "lm", formula = "y~x", se=F)+
  labs(x="Educación en años", 
       y= "Ingresos")



# No linealidades
lm(wage~poly(educ, 2), data = dat)

dat2 %>% 
  ggplot() +
  aes(x= educ, y = wage) +
  geom_jitter() +
  scale_x_continuous(expand = c(0, 0)) +
  geom_smooth(method = "lm", formula = "y~poly(x, 2)", se=F)
  labs(x="Educación en años", 
       y= "Ingresos")