MIT R tutorial

R Orientation

rep(1, 100)

my_object = seq(from=0, to=50, by=2)

plot(my_object)

getwd()

setwd() #use forward slashes even in windows

install.packages("readxl")

library(readxl)

read_excel("data.xlsx")

read.csv("data.csv")

install.packages("foreign")

library(foreign)

read.dta("data.dta")

install.packages("haven")

library(haven)

read_sav("data.sav")

?sample

sample(c(0,1), 100, replace=TRUE)

Import Data

read.csv("data.csv", header=TRUE) # consider there is a header

sample_csv = read.table("data.csv", header=TRUE, sep=",")

 

Assess Data

summary(sample_csv)

nrow(sample_csv)

max(sample_csv)

min(sample_csv)

mean(sample_csv)

median(sample_csv)

unique(sample_csv$E)

sort(sample_csv$E)

length(sample_csv$E)

which(sample_csv$E=0)

length(sample_csv$E=0)

is.matrix(sample_csv)

is.data.frame(sample_csv)

Manipulate Data

as.matrix(sample_csv) # convert data-frame into a matrix

matrix(1, 2, 3)

as.data.frame(a) # convert a matrix into data-frame

t(a) #transpose

a[1,] # call first row

a[-1,] # delete first row

 rbind(a, c(2, 2, 2)) # add a row

 cbind(a, c(2, 3)) # add a column

Sampling, for-loops, and the apply function

states = c("Cairo", "New York", "Nairobi")

sample(x=states, size=10, replace=TRUE)

sample(x=states, size=2, replace=FALSE)

rnorm(n=100)

plot(density(rnorm(n=100000)))

coin = <- c("Heads", "Tails")

 toss = <- c()

 for (i in 1:100) {

    toss[i] <- sample(x=coin, size=1)

}

table(toss)

marital <- c("married", "single")

income <- 1:4

results <- matrix(nrow=100, ncol=3, data=NA)

columns(results) <- c("marital", "income", "state")

head(results)

for (i in 1:100) {

    results[i, 1] <- sample(marital, size=1)

    results[i, 2] <- sample(income, size=1)

    results[i, 3] <- sample(state, size=1)

}

head(results)

apply(X=results, MARGIN=2, FUN=table)

Introduction to the Class lm

height <- c(1, 2, 3)

bodymass <- c(5, 6, 7)

fit <- lm(bodymass ~ height) # Y - X

class(fit)

summary(fit)

names(fit)

plot(height, bodymass) # X - Y

abline(fit, col="red")

Run IV regression using ivreg (from AER package)

# Load the AER package
library(AER)

# General IV regression syntax:
iv_model <- ivreg(
  formula = dependent_var ~ endogenous_var + control_var1 + control_var2 + ... |
            instrument_var + control_var1 + control_var2 + ...,
  data = your_data_frame
)

# Summary of the model
summary(iv_model)
 
dependent_var ~ endogenous_var + controls | instruments + controls

 

# It converts a numeric or character vector into a factor, which is R's way of handling categorical variables 
# (discrete groups, not continuous numbers)
x <- c(1980, 1981, 1982, 1980, 1981)
y <- as.factor(x)

[1] 1980 1981 1982 1980 1981
Levels: 1980 1981 1982