Why Stata Is Better (At Graphing) Than R

Introduction

Both R and Stata are programs with strong data visualization and analysis capabilities. However, Stata’s capabilities as a data visualization program are sometimes under-rated. The intent of the post is to show that Stata can often perform the same graphing task as R, with much simpler, and much more intuitive, command syntax.

This post uses simulated social service agency data clients. In each program, I am going to graph mental health of clients (at Time 2) by program.

Stata

import delimited "clients.csv", encoding(ISO-8859-2) clear // import data
    
graph bar mental_health_t2, /// bar graph of mean outcome
over(program, gap(10)) /// over program
scheme(michigan) /// michigan graph scheme
asyvars // bars that have color

quietly graph export mybarStata.png, width(2000) replace

(8 vars, 491 obs)

Bar Graph in Stata

R

library(readr) # library to import data
    
clients <- read_csv("clients.csv") # import data
    
library(ggplot2)
    
library(michigancolors)

ggplot(clients, # the data that I am using
       aes(x = program, # 'aesthetic' includes x
           y = mental_health_T2, # and y
           fill = program)) + # and fill by program
  stat_summary(fun = mean, # summarizing y 
               geom = "bar") + # with bars
  scale_fill_manual(values = michigancolors()) + # UM colors
  theme_minimal() # minimal theme