1  Getting Started

* Stata: Variables and basic types

* Stata works with datasets, not individual variables
* First, let's create some local macros (temporary variables)

local age = 25
local income = 55000.50
local name "Alice"

* Display values
display "Name: `name', Age: `age'"

* Global macros persist across programs
global project_name "My Analysis"
display "$project_name"

* Create a small dataset
clear
input score
    85
    92
    78
    95
end

* Calculate mean
summarize score
display "Average score: " r(mean)

# R: Variables and basic types

# Numeric types
age <- 25                    # Integer
income <- 55000.50           # Numeric

# Character (string)
name <- "Alice"              # Character

# Logical (boolean)
is_employed <- TRUE          # Logical

# Vectors (R's primary data structure)
scores <- c(85, 92, 78, 95)   # Numeric vector

# List (like Python dictionary)
person <- list(
  name = "Alice",
  age = 25,
  city = "Boston"
)

# Print values
cat("Name:", name, ", Age:", age, "\n")
cat("Average score:", mean(scores), "\n")

* Stata: Arithmetic operations

local a = 10
local b = 3

display "Addition: " `a' + `b'           // 13
display "Subtraction: " `a' - `b'        // 7
display "Multiplication: " `a' * `b'     // 30
display "Division: " `a' / `b'           // 3.333...
display "Integer division: " floor(`a'/`b') // 3
display "Modulo: " mod(`a', `b')        // 1
display "Power: "  `a'^`b'               // 1000

* Comparison operators (return 1 for true, 0 for false)
display "a > b: " (`a' > `b')           // 1
display "a == b: " (`a' == `b')         // 0
display "a != b: " (`a' != `b')         // 1

# R: Arithmetic operations

a <- 10
b <- 3

cat("Addition:", a + b, "\n")          # 13
cat("Subtraction:", a - b, "\n")       # 7
cat("Multiplication:", a * b, "\n")    # 30
cat("Division:", a / b, "\n")          # 3.333...
cat("Integer division:", a %/% b, "\n") # 3
cat("Modulo:", a %% b, "\n")           # 1
cat("Power:", a ^ b, "\n")             # 1000

# Comparison operators
cat("a > b:", a > b, "\n")            # TRUE
cat("a == b:", a == b, "\n")          # FALSE
cat("a != b:", a != b, "\n")          # TRUE

* Stata: Installing and using packages

* Most Stata functionality is built-in
* User-written packages come from SSC (Statistical Software Components)

* Install a package from SSC
ssc install rdrobust      // RD estimation
ssc install outreg2       // Export regression tables
ssc install estout        // More regression output options

* Install from GitHub (using net install)
net install did, from("https://raw.githubusercontent.com/bcallaway11/did/master/stata/") replace

* Update all installed packages
adoupdate, update

* Search for packages
search difference-in-differences

* Check installed packages
ado dir

# R: Installing and loading packages

# Install packages (only need to do once)
install.packages("tidyverse")   # Includes dplyr, ggplot2, tidyr, etc.
install.packages("fixest")      # Fast fixed effects estimation
install.packages("rdrobust")    # Regression discontinuity

# Load packages (do every session)
library(tidyverse)   # Data manipulation & viz
library(fixest)      # Econometrics
library(haven)       # Read Stata/SPSS/SAS files

# Install from GitHub
# install.packages("devtools")
# devtools::install_github("author/package")

# Check package version
packageVersion("tidyverse")

* Stata: Options after the comma act like keyword arguments

* Everything after the comma is an option (named parameter)
import delimited "data.csv", encoding("utf-8") delimiter(";") clear

* summarize with the 'detail' option
summarize income, detail

* regress with the 'robust' option
reg income treatment age, robust

# R: Named arguments work the same way
library(readr)

# Named arguments: delim, locale, na are passed by name
df <- read_delim("data.csv",
                delim = ";",
                locale = locale(encoding = "latin1"),
                na = c("", "NA", "."))

# na.rm = TRUE is a named argument you'll see everywhere
mean(df$income, na.rm = TRUE)

* Stata: Commands are separate (no method chaining)
* Each command is its own line

use "data.dta", clear
describe
summarize income

* Stata achieves "grouping" with the bysort prefix
bysort region: summarize income

# R: The pipe operator %>% chains operations (like method chaining)
library(dplyr)

# The pipe passes the result of each step to the next function
result <- df %>%
  group_by(region) %>%
  summarise(avg_income = mean(income, na.rm = TRUE))

# R 4.1+ also has a native pipe: |>
result <- df |>
  group_by(region) |>
  summarise(avg_income = mean(income, na.rm = TRUE))

* Stata: foreach and forvalues loops

* Loop over a range of numbers
forvalues i = 1/10 {
    display `i'
}

* Loop over a list of items
foreach c in US UK FR {
    display "`c'"
}

* Loop over variables
foreach v of varlist income education age {
    summarize `v'
}

# R: for loops

# Loop over a sequence of numbers
for (i in 1:10) {
  print(i)
}

# Loop over a vector of items
for (country in c("US", "UK", "FR")) {
  print(country)
}

# Loop over column names
for (col in names(df)) {
  print(paste(col, class(df[[col]])))
}

* Stata: Nested functions work the same way

display round(ln(50000), 0.01)
* Step 1: ln(50000)         → natural log = 10.819...
* Step 2: round(..., 0.01)  → round to 2 decimals = 10.82

# R: Nested function calls
incomes <- c(30000, 45000, 52000, 48000, 120000)

# Nested: innermost runs first
result <- round(mean(log(incomes)), 2)

# Or use the pipe to read left-to-right:
result <- incomes %>% log() %>% mean() %>% round(2)

* Stata: Programs are Stata's equivalent of functions

program define compute_tax
    args varname rate
    gen tax = `varname' * `rate'
end

* Call it
compute_tax income 0.25

* Scope: local macros are local to the program
* global macros are visible everywhere (use sparingly)

# R: Function definition and scope

compute_tax <- function(dataframe, rate) {
  dataframe$tax <- dataframe$income * rate
  return(dataframe)
}

# Call the function
df <- compute_tax(df, rate = 0.25)

# R scope: functions look in parent environments for undefined variables
# Same pitfall as Python — always pass values as parameters

* Stata: 1-based indexing
clear
input gdp
100
200
300
end

display gdp[1]   // 100 — first observation
display gdp[2]   // 200 — second observation
display gdp[3]   // 300 — third observation

# R: 1-based indexing
gdp <- c(100, 200, 300)

print(gdp[1])   # 100 — first element
print(gdp[2])   # 200 — second element
print(gdp[3])   # 300 — third element
# gdp[0] returns numeric(0) in R — it doesn't error, just returns empty

* Stata: First script with Gapminder data
* Script: 01_gapminder_analysis.do

* Set working directory (change to your path)
cd "~/Documents/protools_er1"

* Load Gapminder data from URL
import delimited "https://raw.githubusercontent.com/jennybc/gapminder/main/data-raw/08_gap-every-five-years.tsv", clear

* Explore the data
describe
list in 1/5

* Summary statistics
summarize
summarize lifeexp gdppercap pop, detail

* Filter to year 2007
keep if year == 2007

* Calculate average life expectancy by continent
table continent, statistic(mean lifeexp) nformat(%5.1f)

* Alternative: collapse to create summary dataset
preserve
collapse (mean) lifeexp, by(continent)
list
restore

* Create a scatter plot
* -----------------------------------------------------------------------
* twoway: Stata's main graphing command for XY plots
* scatter: plot type - creates a scatter plot of lifeexp (Y) vs gdppercap (X)
* [w=pop]: analytical weights - larger bubbles for countries with larger populations
* msymbol(Oh): marker symbol - hollow circles ("O" = circle, "h" = hollow)
* xscale(log): display X axis on logarithmic scale (useful for GDP data)
* xtitle/ytitle: axis labels
* title: main plot title
* -----------------------------------------------------------------------
twoway (scatter lifeexp gdppercap [w=pop], msymbol(Oh)), xscale(log) xtitle("GDP per capita (log scale)") ytitle("Life expectancy") title("Life Expectancy vs GDP per Capita (2007)")

graph export "gapminder_plot.png", replace

# R: First script with Gapminder data
# Script: 01_gapminder_analysis.R

# Install packages (only need to run once, then comment out)
# install.packages("tidyverse")  # data manipulation and visualization
# install.packages("gapminder")  # example dataset

# Load packages
library(tidyverse)
library(gapminder)

# Explore the data
head(gapminder)
glimpse(gapminder)

# Summary statistics
summary(gapminder)

# Filter to year 2007
gap_2007 <- gapminder %>%
  filter(year == 2007)

# Calculate average life expectancy by continent
life_exp_by_continent <- gap_2007 %>%
  group_by(continent) %>%
  summarize(
    mean_life_exp = mean(lifeExp),
    n_countries = n()
  )

print(life_exp_by_continent)

# Create a scatter plot with ggplot2
ggplot(gap_2007, aes(x = gdpPercap, y = lifeExp,
                      size = pop, color = continent)) +
  geom_point(alpha = 0.7) +
  scale_x_log10(labels = scales::comma) +
  scale_size(range = c(2, 12), guide = "none") +
  labs(
    x = "GDP per capita (log scale)",
    y = "Life expectancy",
    title = "Life Expectancy vs GDP per Capita (2007)",
    color = "Continent"
  ) +
  theme_minimal()

ggsave("gapminder_plot.png", width = 10, height = 6, dpi = 150)