GapMinder_Analysis

Data Description

• Region/Country fields describes the continent and country the Life expectancy is measured
• Additionally, Population, Income, and Life expectance is provided for each country from years 1800 to 2015

Importing/Calling required libraries

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

library(ggplot2)
library(matrixStats)

## 
## Attaching package: 'matrixStats'

## The following object is masked from 'package:dplyr':
## 
##     count

library(scales)

Set working directory and Import raw data

setwd("~/Desktop/EXPL DATA ANALYSIS/EDA REPORT")
gapminder <- read.csv("gapminder.csv")

Summary of data

str(gapminder)

## 'data.frame':    41284 obs. of  6 variables:
##  $ Country   : Factor w/ 197 levels "Afghanistan",..: 1 1 1 1 1 1 1 1 1 1 ...
##  $ 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: Factor w/ 15260 levels "","1,005,328,574",..: 7490 1 1 1 1 1 1 1 1 1 ...
##  $ income    : int  603 603 603 603 603 603 603 603 603 603 ...
##  $ region    : Factor w/ 6 levels "America","East Asia & Pacific",..: 5 5 5 5 5 5 5 5 5 5 ...

Here, we study that Country, population, and income are Factors. Life and income are of num data type while year is of int datatype. We notice that population is a number but due to commas, it is turned a factor. hece we clean the data by removing the commas and creating another field as numeric. ### Study the scope

#Overview of complete Gapminder dataset
summary(gapminder)

##                 Country           Year           life      
##  Afghanistan        :  216   Min.   :1800   Min.   : 1.00  
##  Albania            :  216   1st Qu.:1854   1st Qu.:31.00  
##  Algeria            :  216   Median :1908   Median :35.12  
##  Angola             :  216   Mean   :1907   Mean   :42.88  
##  Antigua and Barbuda:  216   3rd Qu.:1962   3rd Qu.:55.60  
##  Argentina          :  216   Max.   :2015   Max.   :84.10  
##  (Other)            :39988                                 
##    population        income                              region     
##         :25817   Min.   :   142   America                   : 7961  
##  121000 :    6   1st Qu.:   883   East Asia & Pacific       : 6256  
##  14092  :    6   Median :  1450   Europe & Central Asia     :10468  
##  1432000:    6   Mean   :  4571   Middle East & North Africa: 4309  
##  229000 :    6   3rd Qu.:  3483   South Asia                : 1728  
##  2574000:    6   Max.   :182668   Sub-Saharan Africa        :10562  
##  (Other):15437   NA's   :2341

#Number of Unique Countries
length(unique(gapminder$Country))

## [1] 197

#Cleaning of population field. Removing commas and reloading as numeric in a different dataset without NAs.
gapminder$population1 <- as.numeric(gsub(",","",gapminder$population))
gapminder1 <- gapminder %>% filter(!is.na(gapminder$population1))

Looking at the summary, now we know there are 197 countries spread across 6 regions, each having 216 entries for years from 1800 to 2015. The life expectancy ranges from 1 to 84%

Income

hist(gapminder$income)

summary(unique(gapminder$income))

##    Min. 1st Qu.  Median    Mean 3rd Qu.    Max.    NA's 
##     142    2978    6651   12397   14744  182668       1

weightedMedian(gapminder$income, w = gapminder$region, na.rm = TRUE)

## [1] 1182

weightedMean(gapminder$income, w = gapminder$region, na.rm = TRUE)

## [1] 3900.356

d <- ggplot(gapminder1, aes(x = population1, y = income))
d + geom_point()+geom_smooth(method = "lm") + scale_x_continuous(labels = scales::comma) 

## Warning: Removed 823 rows containing non-finite values (stat_smooth).

## Warning: Removed 823 rows containing missing values (geom_point).

Looking at the histogram, a high frequency of income lay in the range below 50000. Further, when we look at the ranges summary, we find that the mean and the median lies in the 2nd quadrant with minimum value of 142 and maximum of 182668
However, things change drastically, when we look at the weighted median and weighted mean based on the region. the Median reduces from 6651 to 1457, while the Mean reduces from 12397 to **4631.
We can say that the income* ranges are different in different regions.
From the plot above (population vs income), we can also find that as the population increases, the income hardly increases for most, while is linear for some. However, if the population is lower always, the income increases exponentially.

###Life Expectancy

Let’s look at the frequencies of the life expectancy and income ranges for the complete data

hist(gapminder$life)

summary(gapminder$life)

##    Min. 1st Qu.  Median    Mean 3rd Qu.    Max. 
##    1.00   31.00   35.12   42.88   55.60   84.10

weightedMedian(gapminder$life, w = gapminder$region, na.rm = TRUE)

## [1] 33.86702

weightedMean(gapminder$life, w = gapminder$region, na.rm = TRUE)

## [1] 41.45024

e <- ggplot(gapminder1, aes(x = life, y = income))
e + geom_point()+geom_smooth(method = "lm") + scale_x_continuous(labels = scales::comma) 

## Warning: Removed 823 rows containing non-finite values (stat_smooth).

## Warning: Removed 823 rows containing missing values (geom_point).

There seems to be a high frequency of life expectancy in the range of 25-35 years
In the next graph, we can see that the life expectancy grows regardless of the income growth. However, as the life expectancy grows past 60, there is a big growth in income.

Population

median(gapminder1$population1)

## [1] 3358089

mean(gapminder1$population1)

## [1] 21187964

weightedMedian(gapminder1$population1, w = gapminder1$Country, na.rm = TRUE)

## [1] 3474657

weightedMean(gapminder1$population1, w = gapminder1$Country, na.rm = TRUE)

## [1] 19078834

The weighted mean is much lower than that of the actual mean when Countries are considered. This mean that there is a big different between population of some countries.

p <- ggplot(gapminder1, aes(reorder(region, population1, FUN=function(x) mean(log10(x))), population1))
p <- p + scale_y_log10()
p + geom_boxplot(outlier.colour="red") + geom_jitter(alpha=1/2)

Comparison between India and China

# Creating subset of India Country
population_india <- gapminder1 %>% filter(Country=='India')
# Creating subset of India Country
population_china <- gapminder1 %>% filter(Country=='China')

ggplot(data= population_india, mapping = aes(y=Year, x=population1, color="India"))  + geom_point(data = gapminder1, aes(color="Others")) + geom_jitter()

ggplot(data= population_china, mapping = aes(y=Year, x=population1, color="China")) + geom_point(data = gapminder1, aes(color="Others")) + geom_jitter()

The population in India and China has risen exponentially after 1950 as compared to other countries.

India and China after 1950

gapminder1 %>% #Filtering data for India
 filter(Country=='India') %>%
summarize('LifeExpected (Mean) in India'=mean(population1),
 'Population (Standard Deviation) in India'=sd(population1))

#Filtering data for India after 1950
pop_ind_after_1950 <- population_india %>% filter(Year > '1950')

b <- ggplot(pop_ind_after_1950, aes(x = population1, y = life))
b + geom_point()+geom_smooth(method = "lm") + scale_x_continuous(labels = scales::comma) + labs(title="India Population vs Life after 1950")

c <- ggplot(pop_ind_after_1950, aes(x = population1, y = income))
c + geom_point()+geom_smooth(method = "lm") + scale_x_continuous(labels = scales::comma) + labs(title="India Population vs Income after 1950") 

gapminder1 %>%filter(Country=='China') %>%#Filtering data for India
summarize('LifeExpected (Mean) in China'=mean(population1),
 'Population (Standard Deviation) in China'=sd(population1))

#Filtering data for India after 1950
pop_chn_after_1950 <- population_china %>% filter(Year > '1950')

b <- ggplot(pop_chn_after_1950, aes(x = population1, y = life))
b + geom_point()+geom_smooth(method = "lm") + scale_x_continuous(labels = scales::comma) + labs(title="China Population vs Life after 1950") 

c <- ggplot(pop_chn_after_1950, aes(x = population1, y = income))
c + geom_point()+geom_smooth(method = "lm") + scale_x_continuous(labels = scales::comma)  + labs(title="China Population vs Income after 1950")

We can see here there income rose for China and India after population crossed 1.1 billion. However, the rise in life expectancy reduced when the population crossed 600 million and flattened between 55 to 65 years.

Income in 2015

#Plotting 2015 data with income vs region
ggplot(gapminder_2015, aes(x = income, y = region)) + facet_wrap(~Year) +
 geom_boxplot(outlier.colour = 'red') +
 ggtitle('Box plots of Income for continents in 2015 across all countries') +
 geom_jitter(position = position_jitter(width = 0.09, height = 0), alpha = 1/10) + scale_x_continuous(labels = scales::comma) 

Here we see that incomes for South Asia, Sub-Saharan Africa, and America is lower than 50000, while other regions go above this range and thus affecting the weighted mean and median.

Life Expectancy in 2015

# Regions with life expectancy mean
ggplot(gapminder_2015, aes(x = region, y = life)) + facet_wrap(~Year) +
 geom_boxplot(outlier.colour = 'red') + # extreme values marked red
 ggtitle('Box plots of life Expectancy continents in 2015') +
 geom_jitter(position = position_jitter(width = 0.09, height = 0), alpha = 1/10)

#List of countries with life expectancy in 2015 arranged in descending order
life_gap_2015 <- gapminder_2015 %>%   group_by(Country) %>%   summarize(life) %>%   arrange(desc(life))
ggplot(data= life_gap_2015, mapping = aes(y=Country, x=life))+   geom_point()+   geom_jitter()+   coord_cartesian(ylim=c(0, 25))

gapminder_2015 %>% # For each continent in 2015
 group_by(region) %>%
 summarize('Standard Deviation'=sd(life),
 'Inter Quartile Range'=IQR(life),
 'Number'=n())

gapminder_2015 %>%
 ggplot(aes(x = life, fill = region)) + facet_wrap(~Year) + # aes = aesthetics
 geom_density(alpha = 0.7) + 
 ggtitle('Density plots of life Expectancy continent in 2015') +
 theme(legend.title = element_text(color = 'Black',size = 14, face = 'bold'),
 legend.background = element_rect(fill = 'gray90', size = 0.5, linetype = 'dashed')) +
 labs(x='Life Expected (Years)', y='Density')

Life Expectancy for America in 2015

Let us focus on a single region and in it a single country. Here, the example considered is the region of America and the country sample is United States.

Calucating the standard deviation for America in 2015:

gapminder_2015 %>% #compute stats
 filter(region=='America') %>%
summarize('LifeExpectancy (Mean) in America for 2015'=mean(life),
 'LifeExpectancy (Standard Deviation) in America for 2015'=sd(life))

America and United States in 2015

# Creating subset of America region
life_America <- gapminder %>% filter(region=='America')
# Creating subset of United States country
life_USA <- gapminder %>%   filter(Country=='United States')

ggplot(data= life_America, mapping = aes(y=Year, x=life, color="America")) + geom_point(data = gapminder, aes(color="Others")) + geom_jitter()

ggplot(data= life_USA, mapping = aes(y=Year, x=life, color="United States")) + geom_point(data = life_America, aes(color="America")) + geom_jitter()