Data Vizualization: The Natural Resource Curse

Introduction

In this demo, I will make an economist style plot from a multiple country regression. For my data I will use country level data from the seminal paper “Natural Resource Abundance and Economic Growth” by Sachs and Warner, which argued that the presence of natural resources in a country caused lower growth.

Setup

I install the package tidverse and I create the pand SAW (Sachs and Warner).

library(tidyverse)
SAW <- read_csv('SAW.csv')
SAW$country = str_remove_all(SAW$country, '\"') 
head(SAW)

Building the Plot

Here’s the most ba

p1 <- ggplot(SAW, mapping = aes(x=sxp, y=gea7090)) + geom_point(size=2) + scale_x_continuous(limits = c(0,.55))
p1

Adding a trendline to the data makes the relationship more clear. Of course, many other factors influence GDP growth, which I did not control for, so this regression says little about real causal relationships.

p2 <- p1 + geom_smooth(method = "lm", formula = y ~ x, level = .95, fullrange=TRUE) + theme_bw()
p2

Let’s label a few of these anonymous points.

label_these <- c("JAPAN","MALAYSIA","ICELAND","FIJI","GAMIA","MAURITANIA","NICARAGUA","GUYANA","VENEZUELA","ARGENTINA","INDONESIA")
  
p3 <- p2 + geom_text(data=filter(SAW, country %in% label_these), mapping = aes(label=country),  check_overlap = FALSE, nudge_x = .07)
p3

Looking good but it could use some style. Let’s apply the econmist style to the plot.

library(ggthemes)
p3 + theme_economist() + labs(title="Natural Resources and Growth", x="Share of Primary Products in Exports", y="Average Yearly GDP Growth 1970-1990", color="Log of GDP in 1970")

With this lovely plot, my demo is complete.