EDA Assignment Summary

In this exploratory data analysis assignment I have used the Gapminder Dataset for analysis. After starting with a brief data inspection and exploratory plotting, I set up three specific questions, shown in section 1 below, for my analysis in this project. Thre reset of the presentation and anlysis of the project is as follows.

I first present a systematic data description which includes description of each variables, total number of records and missing values etc. In Data exploration section I have first presented regionwise (subcontinents) discriptive statistics and visulization (boxplots, histograms, and density plots for different) of the different variables present in the data. The analysis results/interpretations are shown below.

The analysis of the Life Expectancy vs Per Capita Income plot did not show any conclusive trend, however, plotting percapita income on log scale shows a nearly linear trend with life expectancy with quite a bit of spread. This is ecpected that increase in percapita income will increase life expectancy.
The life expectancy trends over the years show a very low and flat expectancy until early parts of the 20th century in all the continents (may be because of lack of health care) and then shown increasing treands in all continents. The increasing trends are set early in high percapita regions, like, US and Europe, and late in poorer regions like middle east/north africa and Asia pacific. There trends are obvious and are in line with the expectations.
Similar analysis for United States shows a clearer tresnd for life expectancy whrere above certain level (~$8000) it increases sharply and the slows down beyond certain limit as income cann’t control life beyond certain limit.Similarly, life expectancy in US started increasing very early (before 1900) and is showing a contineous increase over time probably due to improvement in health care, which is also strongly related to income as well.

1. Questions

Q1) Analyze trends in Life Expectancy (i.e. life) with Per Capita Income (i.e. Income) for the entire dataset.

Q2) Analyse trends in Life Expectancy (i.e. life) over the years for different subcontinents (i.e. regions).

Q3) Repeat Q1 and Q2 analysis for United Staes.

2. Data description

library(readr)
gapminder <- read_csv("C:/Users/rs443/Dropbox/My courses/R-Courses/ANLY 506/Final Report/gapminder.csv")

## Parsed with column specification:
## cols(
##   Country = col_character(),
##   Year = col_integer(),
##   life = col_double(),
##   population = col_number(),
##   income = col_double(),
##   region = col_character()
## )

The gapminder dataset is a comma-separated value (.CSV) file with 41284 records.The data frame has six features/variables

Country (a categorical variable), a factor with 197 levels
region (a categorical variable), a factor with 6 levels represent different subcontinental regions in the world

Year (a categorical variable), a factor with 216 levels

life (a continuous variable)- denote Life Expectancy
population(a discrete variable)
income (a continuous variable)- represents Per Capita Income

nrow(gapminder)

## [1] 41284

ncol(gapminder)

## [1] 6

str(gapminder)

## Classes 'tbl_df', 'tbl' and 'data.frame':    41284 obs. of  6 variables:
##  $ Country   : chr  "Afghanistan" "Afghanistan" "Afghanistan" "Afghanistan" ...
##  $ Year      : int  1800 1801 1802 1803 1804 1805 1806 1807 1808 1809 ...
##  $ life      : num  28.2 28.2 28.2 28.2 28.2 ...
##  $ population: num  3280000 NA NA NA NA NA NA NA NA NA ...
##  $ income    : num  603 603 603 603 603 603 603 603 603 603 ...
##  $ region    : chr  "South Asia" "South Asia" "South Asia" "South Asia" ...
##  - attr(*, "spec")=List of 2
##   ..$ cols   :List of 6
##   .. ..$ Country   : list()
##   .. .. ..- attr(*, "class")= chr  "collector_character" "collector"
##   .. ..$ Year      : list()
##   .. .. ..- attr(*, "class")= chr  "collector_integer" "collector"
##   .. ..$ life      : list()
##   .. .. ..- attr(*, "class")= chr  "collector_double" "collector"
##   .. ..$ population: list()
##   .. .. ..- attr(*, "class")= chr  "collector_number" "collector"
##   .. ..$ income    : list()
##   .. .. ..- attr(*, "class")= chr  "collector_double" "collector"
##   .. ..$ region    : list()
##   .. .. ..- attr(*, "class")= chr  "collector_character" "collector"
##   ..$ default: list()
##   .. ..- attr(*, "class")= chr  "collector_guess" "collector"
##   ..- attr(*, "class")= chr "col_spec"

head(gapminder)

## # A tibble: 6 x 6
##   Country      Year  life population income region    
##   <chr>       <int> <dbl>      <dbl>  <dbl> <chr>     
## 1 Afghanistan  1800  28.2    3280000    603 South Asia
## 2 Afghanistan  1801  28.2         NA    603 South Asia
## 3 Afghanistan  1802  28.2         NA    603 South Asia
## 4 Afghanistan  1803  28.2         NA    603 South Asia
## 5 Afghanistan  1804  28.2         NA    603 South Asia
## 6 Afghanistan  1805  28.2         NA    603 South Asia

tail(gapminder)

## # A tibble: 6 x 6
##   Country  Year  life population income region               
##   <chr>   <int> <dbl>      <dbl>  <dbl> <chr>                
## 1 Åland    1992  80.8      24834     NA Europe & Central Asia
## 2 Åland    1993  81.8      24950     NA Europe & Central Asia
## 3 Åland    1994  80.6      25066     NA Europe & Central Asia
## 4 Åland    1995  79.9      25183     NA Europe & Central Asia
## 5 Åland    1996  80        25301     NA Europe & Central Asia
## 6 Åland    1997  80.1      25419     NA Europe & Central Asia

3 Data exploration

3.1 Exploratory regional data analysis

3.1.1 Data summary statistics

a) Mean and Median of life expectancy Region wise

aggregate(life ~ region, gapminder, median)

##                       region     life
## 1                    America 35.37370
## 2        East Asia & Pacific 34.00000
## 3      Europe & Central Asia 41.74110
## 4 Middle East & North Africa 32.30000
## 5                 South Asia 32.64700
## 6         Sub-Saharan Africa 32.40338

aggregate(life ~ region, gapminder, mean)

##                       region     life
## 1                    America 44.54065
## 2        East Asia & Pacific 41.76041
## 3      Europe & Central Asia 48.79419
## 4 Middle East & North Africa 41.55366
## 5                 South Asia 37.42302
## 6         Sub-Saharan Africa 37.88242

b) Five Num Summary Life Expectancy Region wise

library(dplyr)

## 
## Attaching package: 'dplyr'

## The following objects are masked from 'package:stats':
## 
##     filter, lag

## The following objects are masked from 'package:base':
## 
##     intersect, setdiff, setequal, union

SouthAsia=filter(gapminder, region=="South Asia")
fivenum(SouthAsia$life)

## [1]  7.045063 27.345157 32.647000 44.374110 79.500000

fivenum(SouthAsia$population)

## [1]      42378     993915   15661841   69328216 1311050527

fivenum(SouthAsia$income)

## [1]   603.0   840.0  1020.0  1286.5 14408.0

EastAsiaPacific = filter(gapminder,region=="East Asia & Pacific")
fivenum(EastAsiaPacific$life)

## [1]  1.00000 28.94900 34.00000 57.49076 83.50000

fivenum(EastAsiaPacific$population)

## [1]       1548     115956    1241248   18196783 1376048943

fivenum(EastAsiaPacific$income)

## [1]    363.0    873.0   1153.5   2557.5 134864.0

EuropeCentralAsia=filter(gapminder,region=="Europe & Central Asia")
fivenum(EuropeCentralAsia$life)

## [1]  3.98908 35.60000 41.74110 66.60000 84.10000

fivenum(EuropeCentralAsia$population)

## [1]      9584   1960597   4750396  10592125 148435811

fivenum(EuropeCentralAsia$income)

## [1]   393  1427  2735  7387 96245

MiddleEastNorthAfrica = filter(gapminder,region=="Middle East & North Africa")
fivenum(MiddleEastNorthAfrica$life)

## [1]  1.50000 30.71320 32.30000 53.70766 82.40000

fivenum(MiddleEastNorthAfrica$population)

## [1]     2788   548618  2615753  9499387 91508084

fivenum(MiddleEastNorthAfrica$income)

## [1]    715.0   1082.0   1537.5   3939.0 182668.0

America =filter(gapminder,region=="America")
fivenum(America$life)

## [1]  9.690052 32.124000 35.373700 61.048460 81.700000

fivenum(America$population)

## [1]      9899.0    236909.5   2144973.0   8510081.5 321773631.0

fivenum(America$income)

## [1]   529.0  1277.5  2214.5  5461.5 53354.0

SubSaharanAfrica = filter(gapminder,region=="Sub-Saharan Africa")
fivenum(SubSaharanAfrica$life)

## [1]  4.00000 30.43680 32.40338 46.00000 79.64600

fivenum(SubSaharanAfrica$population)

## [1]      8219    821457   3019768   8719290 182201962

fivenum(SubSaharanAfrica$income)

## [1]   142   596   827  1323 40143

3.1.2. Raw Data Visulization

a) Boxplots

Life expectancy comparision in different region (subcontenets)

library(ggplot2)
ggplot(gapminder, aes(x = region, y = life)) +  geom_boxplot(outlier.colour = "red") +  geom_jitter(position = position_jitter(width = 0.1, height = 0), alpha = 1/4)

Population comparision in different region (subcontenets)

ggplot(gapminder, aes(x = region, y = population)) +  geom_boxplot(outlier.colour = "red") +  geom_jitter(position = position_jitter(width = 0.1, height = 0), alpha = 1/4)

## Warning: Removed 25817 rows containing non-finite values (stat_boxplot).

## Warning: Removed 25817 rows containing missing values (geom_point).

Income comparision in different region (subcontenets)

ggplot(gapminder, aes(x = region, y = income)) +  geom_boxplot(outlier.colour = "red") +  geom_jitter(position = position_jitter(width = 0.1, height = 0), alpha = 1/4)

## Warning: Removed 2341 rows containing non-finite values (stat_boxplot).

## Warning: Removed 2341 rows containing missing values (geom_point).

b) Histograms

Life Expectancy by region

library(ggplot2)
par(mfrow=c(2,3))
hist(America$life)
hist(EastAsiaPacific$life)
hist(EuropeCentralAsia$life)
hist(MiddleEastNorthAfrica$life)
hist(SouthAsia$life)
hist(SubSaharanAfrica$life)

c) Density plots

Density Plot of life expectancy by region

ggplot(gapminder,aes(life))+geom_density(aes(color=region))

3.1.3. Analysis of Life Expectancy vs Per capita income for different subcontinents

plot(gapminder$income, gapminder$life, main = "Life Expectancy vs Per Capita Income" )

plot(log(gapminder$income), gapminder$life, main = "Life Expectancy vs log(Per Capita Income)")

3.1.4 Life expectancy trends over the years for different regions/subcontinents

par(mfrow=c(2,3))
plot(SouthAsia$Year, SouthAsia$life, main = "SouthAsia")
plot(EastAsiaPacific$Year, EastAsiaPacific$life, main = "Asia Pacific")
plot(EuropeCentralAsia$Year, EuropeCentralAsia$life, main = "Europe and Central Asia")
plot(MiddleEastNorthAfrica$Year, MiddleEastNorthAfrica$life, main = "Middle East and North Africa")
plot(America$Year, America$life, main = "America")
plot(SubSaharanAfrica$Year, SubSaharanAfrica$life, main = "SubSahara and Africa")

3.2 Exploratory data analysis for a single country (United States)

gapminderUS=filter(gapminder,Country=="United States")

Data summary and statistics

Five-number summaries for different variables for the US

fivenum(gapminderUS$life)

## [1] 31.000 39.410 50.550 69.938 79.100

fivenum(gapminderUS$population)

## [1]   6801854 170796378 218963561 266275528 321773631

fivenum(gapminderUS$income)

## [1]  2115.0  3505.0  7875.0 19231.5 53354.0

Raw data visulization

a) Boxplots for different variables for the US

par(mfrow=c(1,3))
boxplot(gapminderUS$life, main = "Life expectancy")
boxplot(gapminderUS$population, main = "Population")
boxplot(gapminderUS$income, main = "Income")

b) Histograms for different variable for the US

par(mfrow=c(1,3))
hist(gapminderUS$life,col = "green")
hist(gapminderUS$population,col = "blue")
hist(gapminderUS$income,col="magenta")

c) Density Plots for different variables for the US

par(mfrow=c(1,2))
plot(density(gapminderUS$life))
plot(density(gapminderUS$income))

3.2.3. Analysis of ‘Life Expectancy’ vs ‘Per Capita Income’ for the United States

par(mfrow=c(1,2))
plot(gapminderUS$income, gapminderUS$life, main = "Life expectancy vs Income")
plot(log(gapminderUS$income), gapminderUS$life, main = "Life expectancy vs Income")

3.2.3. Analysis of ‘Life Expectancy’, ‘Population’ and ‘Per Capita Income’ variation over the years for the United States

par(mfrow=c(1,3))
plot(gapminderUS$Year, gapminderUS$life, main = "Life expectancy vs Year")     
plot(gapminderUS$Year, gapminderUS$population, main = "Population vs Year")     
plot(gapminderUS$Year, gapminderUS$income, main = "Income vs Year")

Exploratory Data Analysis Project- Gapminder Dataset

Anjali Rawat

February 18, 2019