Pick Variables

The list of columns that we use in this project is Country.name, year, Life.Ladder, Log.GDP.per.capital, Social.support, Healthy.life.expectancy.at.birth, Freedom.to.make.life.choices,Generosity , and Perceptions.of.corruption.

Drop Unused Columns

As illustrated from str() result, we need to drop several columns in data1 and data2. Then, check the result with colnames().

# Drop columns in data1
data1 <- data1[, -c(10,11)]
colnames(data1)

## [1] "Country.name"                     "year"                            
## [3] "Life.Ladder"                      "Log.GDP.per.capita"              
## [5] "Social.support"                   "Healthy.life.expectancy.at.birth"
## [7] "Freedom.to.make.life.choices"     "Generosity"                      
## [9] "Perceptions.of.corruption"

# Drop columns in data2
data2 <- data2[, -c(4:6)]
data2 <- data2[, -c(10:17)]
colnames(data2)

## [1] "Country.name"                 "Regional.indicator"          
## [3] "Ladder.score"                 "Logged.GDP.per.capita"       
## [5] "Social.support"               "Healthy.life.expectancy"     
## [7] "Freedom.to.make.life.choices" "Generosity"                  
## [9] "Perceptions.of.corruption"

Create a New Variable

Since data2 covers data for 2021, we need to create the new column year in data frame data2 and fill its values with 2021. Then, check the result by using colnames().

data2$year <- 2021
colnames(data2)

##  [1] "Country.name"                 "Regional.indicator"          
##  [3] "Ladder.score"                 "Logged.GDP.per.capita"       
##  [5] "Social.support"               "Healthy.life.expectancy"     
##  [7] "Freedom.to.make.life.choices" "Generosity"                  
##  [9] "Perceptions.of.corruption"    "year"

Change Column Names

We found the names in both data frames are not consistent and too long. To make it better and be able to merge both data frames, perhaps we can change the name.

# Changing column names of data1
colnames(data1) <- c("Country", "year", "Score","Log.GDP", "Social.support", "Healthy", "Freedom","Generosity" , "Corruption")

# Changing column names of data2
colnames(data2) <- c("Country" ,"Continent", "Score", "Log.GDP", "Social.support", "Healthy", "Freedom","Generosity" , "Corruption", "year")

Join Both Data Frames

This section contains many steps.

• Create a new data frame called continent, which containing the column Country and Continent from data2.
• Merge continent into data1 by using merge.
• Utilize bind_rows() from library dplyr to bind data1 and data2, and assign it as new data frame called happy.
• Check the data frame happy by applying str().

# Select only columns `Country `and `Continent` from `data2` and assigned as `continent`
continent <- data2[, c("Country", "Continent") ]

# Join `continent` by Country into `data1`
data1 <- merge(data1, continent, by = "Country")

# Stacking `data1` and `data2` and assigned as `happy`
happy <- bind_rows(data1, data2) 

str(happy)

## 'data.frame':    2035 obs. of  10 variables:
##  $ Country       : chr  "Afghanistan" "Afghanistan" "Afghanistan" "Afghanistan" ...
##  $ year          : num  2014 2015 2016 2017 2008 ...
##  $ Score         : num  3.13 3.98 4.22 2.66 3.72 ...
##  $ Log.GDP       : num  7.72 7.7 7.7 7.7 7.37 ...
##  $ Social.support: num  0.526 0.529 0.559 0.491 0.451 0.521 0.521 0.484 0.42 0.508 ...
##  $ Healthy       : num  52.9 53.2 53 52.8 50.8 ...
##  $ Freedom       : num  0.509 0.389 0.523 0.427 0.718 0.496 0.531 0.578 0.394 0.374 ...
##  $ Generosity    : num  0.104 0.08 0.042 -0.121 0.168 0.162 0.236 0.061 -0.108 -0.094 ...
##  $ Corruption    : num  0.871 0.881 0.793 0.954 0.882 0.731 0.776 0.823 0.924 0.928 ...
##  $ Continent     : chr  "South Asia" "South Asia" "South Asia" "South Asia" ...

Treatment for Missing Value.

Now, happy has 2035 rows and 10 variables. Let’s observe missing values for each column by using colSums() and is.na().

colSums(is.na(happy))

##        Country           year          Score        Log.GDP Social.support 
##              0              0              0             24              9 
##        Healthy        Freedom     Generosity     Corruption      Continent 
##             51             30             76            104              0

Check the percentage of missing values for each column. It seems that the proportion of missing values for each column is still acceptable for us to keep the data.

missing.values.col <- round(as.data.frame(apply(happy, 2, function(col)sum(is.na(col))*100/length(col))),2)
missing.values.col

##                apply(happy, 2, function(col) sum(is.na(col)) * 100/length(col))
## Country                                                                    0.00
## year                                                                       0.00
## Score                                                                      0.00
## Log.GDP                                                                    1.18
## Social.support                                                             0.44
## Healthy                                                                    2.51
## Freedom                                                                    1.47
## Generosity                                                                 3.73
## Corruption                                                                 5.11
## Continent                                                                  0.00

Check Duplication

Check duplicate data in row by using duplicated. Great no duplicates!

happy[duplicated(happy),]

##  [1] Country        year           Score          Log.GDP        Social.support
##  [6] Healthy        Freedom        Generosity     Corruption     Continent     
## <0 rows> (or 0-length row.names)

Datatype Transformation

Transform Country , year and Continent as factor by applying as.factor(). Sneak peak with str() for the end result of data preparation processes.

happy[, c("Country", "Continent", "year")] <- lapply(happy[, c("Country", "Continent", "year")], as.factor)
str(happy)

## 'data.frame':    2035 obs. of  10 variables:
##  $ Country       : Factor w/ 149 levels "Afghanistan",..: 1 1 1 1 1 1 1 1 1 1 ...
##  $ year          : Factor w/ 17 levels "2005","2006",..: 10 11 12 13 4 7 8 9 15 14 ...
##  $ Score         : num  3.13 3.98 4.22 2.66 3.72 ...
##  $ Log.GDP       : num  7.72 7.7 7.7 7.7 7.37 ...
##  $ Social.support: num  0.526 0.529 0.559 0.491 0.451 0.521 0.521 0.484 0.42 0.508 ...
##  $ Healthy       : num  52.9 53.2 53 52.8 50.8 ...
##  $ Freedom       : num  0.509 0.389 0.523 0.427 0.718 0.496 0.531 0.578 0.394 0.374 ...
##  $ Generosity    : num  0.104 0.08 0.042 -0.121 0.168 0.162 0.236 0.061 -0.108 -0.094 ...
##  $ Corruption    : num  0.871 0.881 0.793 0.954 0.882 0.731 0.776 0.823 0.924 0.928 ...
##  $ Continent     : Factor w/ 10 levels "Central and Eastern Europe",..: 7 7 7 7 7 7 7 7 7 7 ...

Question 1

Discover countries and continents with the highest and lowest life satisfaction score in 2021.

In this section, we will create two plots and one table to answer the question.

1. Bar plot to illustrate the country with the highest and the lowest life satisfaction score.
2. Table of the mean values for each continent to show the continent with highest and the lowest score and also distribution plot (ridgeline) to visualize the distribution of life satisfaction score in each continent.

Question 1 (a)

Create a data frame called happy.2021 to filter data in 2021. Then make a new data frame containing the list of 10 the happiest and the least happy country called happy.country.head and happy.country.tail, respectively. Merge both data frames into head.tail.

# Filter the data frame with `year` is 2021 and assigned as `happy.2021`. Order by `Score` column.
happy.2021 <- happy[happy$year == 2021,]
happy.2021 <- happy.2021[order(-happy.2021$Score),]

# Create data frame of 10 the happiest countries called `happy.country.head` and the least happy countries called `happy.country.tail`. 
happy.country.head <- head(happy.2021, 10)
happy.country.tail <- tail(happy.2021, 10)
head.tail <- rbind(happy.country.head, happy.country.tail)

Construct the plot to visualize the question by using ggplot. Note : geom_chicklet behaves similar to ggplot2::geom_col() but gives an option to turn sharp-edged bars into rounded rectangles.

ggplot(head.tail, aes(x = reorder(Country, Score))) + 
  
  geom_chicklet(aes(y = 10, fill = 5.7), width = 0.9, radius = grid::unit(10, "pt")) +
  geom_chicklet(aes(y = Score, fill = Score), width = 0.9, radius = grid::unit(10, "pt")) +
  coord_flip() +
  
  geom_text(aes(y = Score), label = round(head.tail$Score,2), nudge_y = -0.4, size = 4, color='white') + 
  scale_y_continuous(expand = c(0, 0.1), position = "right", limits = c(0, 10)) +
  scale_fill_gradient2(low = "#C3B236", high = '#224709', mid = 'white', midpoint = 6) + 
  
  theme_ipsum(grid = '') +
  labs(
    x = NULL, y = "Best score = 10", fill = NULL,
    title = "Countries with The Highest and \nThe Lowest Life Satisfaction Score in 2021",
    subtitle = "Total countries in dataset is 149. However, the plot only shows\n20 countries which has the highest and the lowest score",
    caption = "Source : The World Happiness Report 2021") +
  theme(plot.title = element_text(size= 20, color = 'black', face ='bold'),
        plot.subtitle = element_text(size= 12, color = 'black'),
        plot.caption = element_text(size= 12, color = 'black'),
        axis.title.x = element_text(size= 12, color = '#555955'),
        axis.text.y = element_text(color = "black", size = 14),
        axis.text.x = element_blank(),
        legend.position = "none") 

Insight: Based on the above plot, nordic countries such as Finland, Denmark, Switzerland and Iceland have the highest scores with an average score above 7. In the same year, the lowest score corresponds to Afghanistan (below 3), followed by Zimbabwe and Rwanda.

Question 1 (b)

Aggregate the happy.2021 data frame by continent and get its mean score value. Assign the result to happy.2021.con and order by Score column.

library(moments)
happy.2021.con <- aggregate(Score ~ Continent, happy.2021, mean)
happy.2021.con$Mean.2021 <- mean(happy.2021.con$Score)
happy.2021.con[order(-happy.2021.con$Score),]

##                             Continent    Score Mean.2021
## 6               North America and ANZ 7.128500   5.67772
## 10                     Western Europe 6.914905   5.67772
## 1          Central and Eastern Europe 5.984765   5.67772
## 4         Latin America and Caribbean 5.908050   5.67772
## 3                           East Asia 5.810333   5.67772
## 2  Commonwealth of Independent States 5.467000   5.67772
## 8                      Southeast Asia 5.407556   5.67772
## 5        Middle East and North Africa 5.219765   5.67772
## 9                  Sub-Saharan Africa 4.494472   5.67772
## 7                          South Asia 4.441857   5.67772

Insight:

• Refer to the table, North America and ANZ (7.12) and Western Europe (6.91) are continents with a high life satisfaction score. On another side, South Asia (4.44) and Sub-Saharan Africa (4.44) have a low score.
• Those four countries’ scores are located far away from the mean value of 5.67. Whereas the other continents score remains close to the mean value.

Since distribution is as important as the average or mean method to compare the life satisfaction score. Now, we will construct ridgeline plot to illustrate the distribution of the score on each continent.

ggplot(happy.2021, aes(x = Score, y = Continent, fill = stat(x))) +
  geom_density_ridges_gradient(scale = 2.5, size = 0.7, rel_min_height = 0.02)+
  scale_fill_gradientn(colours =  wes_palette("Moonrise2", 10, type = "continuous"), name = "Score Level")+
  theme_ipsum() +
  labs(
    x = "Life Satisfaction Score", y = "", fill = "",
    title = "Distribution Plot of Life Satisfaction\nScore Based on Continent in 2021",
    caption = "Source : The World Happiness Report 2021") +
  theme(plot.title = element_text(size= 20, color = 'black', face ='bold'),
        plot.caption = element_text(size= 12, color = 'black'),
        axis.title.x = element_text(size= 13, color = '#555955'),
        axis.text.y = element_text(color = "black", size = 14),
        axis.text.x = element_text(size= 12, color = 'black'),
        legend.position = "bottom") 

## Picking joint bandwidth of 0.281

Insight :

• As we expected, scores in Western Europe and North America are mostly distributed approximately above 7. North America is normally distributed, while Western Europe has a left skew distribution.
• Scores in Sub-Saharan Africa, the Middle East and North Africa are mostly dispersed below 5. It also implies that those continents have been undermined by other continents.
• The rest of the continents’ scores are distributed between 5 and 6.5.

Question 2

Discover the correlation between each factor in the life satisfaction level.

In this section, we will create three plots containing the correlation of factors that affect life satisfaction. The first two bubble plots are representative plots to observe a correlation between a few factors to the score. The last plot is a correlation matrix, which serves as the main plot to gather a conclusion in this section.

1. A bubble plot describing a correlation between life satisfaction score, life expectancy and Log GDP per capita
2. A bubble plot describing a correlation between life satisfaction score, freedom and life expectancy.
3. Correlation matrix chart.

For better visualization, we will create a new variable called Cat.Continent inside the happy.2021 dataframe to divide the continent into three continents.

• Western Europe/North America and ANZ
• Sub-Saharan African/South Asia
• Other continents

# Create a function to divide Continent into three parts.
convert_continent <- function(y){ 
    if( y == "Western Europe" | y == "North America and ANZ"){
      y <- "Western Europe/North America"
    }else 
      if(y == "Sub-Saharan Africa" | y == "South Asia"){
      y <- "Sub-Saharan Africa/South Asia" 
    }else{
      y <- "Others" 
    }  
}

# Create a variable called Cat.Continent by using `sapply` and transform data type into factor. Then, check dataframe by using `str()`
happy.2021$Cat.Continent <- sapply(X = happy.2021$Continent, FUN = convert_continent) 
happy.2021$Cat.Continent <- as.factor(happy.2021$Cat.Continent)
str(happy.2021)

## 'data.frame':    149 obs. of  11 variables:
##  $ Country       : Factor w/ 149 levels "Afghanistan",..: 41 34 129 55 97 104 128 79 98 7 ...
##  $ year          : Factor w/ 17 levels "2005","2006",..: 17 17 17 17 17 17 17 17 17 17 ...
##  $ Score         : num  7.84 7.62 7.57 7.55 7.46 ...
##  $ Log.GDP       : num  10.8 10.9 11.1 10.9 10.9 ...
##  $ Social.support: num  0.954 0.954 0.942 0.983 0.942 0.954 0.934 0.908 0.948 0.934 ...
##  $ Healthy       : num  72 72.7 74.4 73 72.4 73.3 72.7 72.6 73.4 73.3 ...
##  $ Freedom       : num  0.949 0.946 0.919 0.955 0.913 0.96 0.945 0.907 0.929 0.908 ...
##  $ Generosity    : num  -0.098 0.03 0.025 0.16 0.175 0.093 0.086 -0.034 0.134 0.042 ...
##  $ Corruption    : num  0.186 0.179 0.292 0.673 0.338 0.27 0.237 0.386 0.242 0.481 ...
##  $ Continent     : Factor w/ 10 levels "Central and Eastern Europe",..: 10 10 10 10 10 10 10 10 6 10 ...
##  $ Cat.Continent : Factor w/ 3 levels "Others","Sub-Saharan Africa/South Asia",..: 3 3 3 3 3 3 3 3 3 3 ...

Question 2 (a and b)

Now, create two bubble plots in one page to describe the correlation as requested by using ggplot and geom_point. Then, utilize grid.arrange to show both plots into one page.

# Create ggplot by using geom_point to create bubble plot to visualize the correlation between life expentencies, score and GDP. Assign as `plot3`
plot3 <-ggplot(happy.2021, mapping =aes(x = Healthy, y= Score, size= Log.GDP,  fill= Cat.Continent)) +
    geom_point(alpha=0.7, shape=21, color="black") +
    scale_size(range = c(.1, 10), name="Log GDP per Capita") +
    scale_fill_manual(values = wes_palette(3, name = "Cavalcanti1"), name = "Continent")  +
    labs(x = "Life Expectencies", y = "Life Satisfaction Score",
         title = "(A) Correlation between Life Satisfaction Score, \nLife Expectencies, and GDP")+
    theme(plot.title = element_text(size= 10, color = 'black', face ='bold'),
         legend.position = "right") +
    theme_ipsum() 

# Create ggplot by using geom_point to create bubble plot to visualize the correlation between corruption, score and freedom. Assign as `plot4`
plot4 <- ggplot(happy.2021, mapping =aes(x = Corruption, y= Score, size= Freedom,  fill= Cat.Continent)) +
    geom_point(alpha=0.7, shape=21, color="black") +
    scale_size(range = c(.1, 10), name="Freedom") +
    scale_fill_manual(values = wes_palette(3, name = "Cavalcanti1"), name = "Continent") +
    labs(x = "Corruption", y = "Life Satisfaction Score",
         title = "(B) Correlation between Life Satisfaction Score, \nCorruption, and Freedom",
         caption =  "The World Happiness Report 2021") +
    theme(plot.title = element_text(size= 10, color = 'black', face ='bold'),
         plot.caption =  element_text(size= 8, color = 'black'),
         legend.position = "right") +
    theme_ipsum() 

# Apply grid.arrange to print both plot into one page. 
grid.arrange(plot3, plot4, nrow=2)

Insight:

• Based on plot (A), there is a strong correlation between the life satisfaction score and average life expectancy. So that, continents which have a high score, indeed also have a high life expectancy compare to continents with low scores.

• The correlation between wealth (GDP) and the score tends to have similar behaviour with the life expectancy correlation. Small bubbles (dark green colour) are mostly clustered in a low score, while big bubbles appear in a high score. To conclude, people living in a country that has a high life expectancy and GDP tends to be ‘happier’ than people living in a country with a low life expentency and GDP.

• Refer to plot (B), the corruption level has a negative correlation with the score. It seems that the condition is only valid for continents with having low and high scores, as illustrated in green and grey bubbles. Since the “Other” continent (yellow), their level of corruption shows a high value, which should be at the middle level.

• There is no correlation between freedom and score of life satisfaction.

Question 2 (c)

First, create a new data frame called global.corr containing numeric columns. Then, plot correlation matrix by using ggcorr

# Selecting variable called `global.corr` for the correlation matrix
global.corr <- select(happy.2021, -c("Country","year","Continent", "Cat.Continent"))

# Plot the correlation matrix
ggcorr(global.corr, 
       method = c("everything", "pearson"), 
       size = 3, hjust = 0.1, color='black', angle=90,
       low = "#C3B236", high = '#224709', mid = 'White',
       label = TRUE, label_size = 3,
       layout.exp = 3) +
labs(title = 'Correlation Matrix',
    caption = "The World Happiness Report 2021")+
theme_ipsum() +
theme(plot.title = element_text(size=18),
      plot.subtitle = element_text(size = 12),
       plot.caption = element_text(size = 12),
      legend.text = element_text(size = 12))

Insight: Referring to the correlation matrix, the life satisfaction score has a strong positive correlation with life expectancies, social support, wealth or GDP and Freedom. While corruption has a negative connection with score. It is interesting to note that corruption and generosity have a small or almost no connection with the other factors. Therefore, higher GPD, social support and life expectancy and lower corruption lead to ‘happier’ countries.

Question 3

Discover the insight of score’s comparison or changes in 2011 and 2021 (ten years score differences)

In this section, we will answer the question based on the country and also the continent. Therefore, we’ll create two A dumbbell plot

1. A dumbbell plot to compare the score in 2011 and 2021 and only show 20 countries that have the biggest life satisfaction score movement.
   
2. A dumbbell plot to compare the score in 2011 and 2021 for each continent.

Question 3 (a)

Let’s create a new data frame called mover from happy containing variables year, Country,Score, Continent, class.score. Check date frame with str().

mover <- happy[, c("year", "Country", "Score", "Continent")]
mover <- mover[mover$year %in% c(2011, 2021),]
str(mover)

## 'data.frame':    285 obs. of  4 variables:
##  $ year     : Factor w/ 17 levels "2005","2006",..: 7 7 7 7 7 7 7 7 7 7 ...
##  $ Country  : Factor w/ 149 levels "Afghanistan",..: 1 2 3 4 5 6 7 8 9 10 ...
##  $ Score    : num  3.83 5.87 5.32 6.78 4.26 ...
##  $ Continent: Factor w/ 10 levels "Central and Eastern Europe",..: 7 1 5 4 2 6 10 2 5 7 ...

Create the data frame mean.sd to get the mean value for each year.

mean.sd <- aggregate(Score ~ year, mover, mean)
mean.sd

##   year    Score
## 1 2011 5.444971
## 2 2021 5.532839

This section is relatively a bit longer. Bottom line, we want to create a data frame mover1 containing 20 countries with the biggest score movement. Then, also create two separate data frames containing values for 2011 and 2021. The step is written in the chunk. Note: the step could be reduced if we use a library(dplyr)

# Use pivot_wider to move score in 2011 and 2021 into a column. 
mover <-pivot_wider(data= mover,
                    names_from = "year", 
                    values_from = "Score")

# Change the name of columns
colnames(mover) <- c("Country", "Continent", "year.2011", "year.2021")

# Create a new variable called `gap` containing an absolute gap values and normal values between both year. 
mover$gap.nor <- mover$year.2021 - mover$year.2011
mover$gap <- abs(mover$year.2021 - mover$year.2011)

# Order based on the gap columns 
mover <- mover[order(-mover$gap),]

# Delete missing values and assign as `mover.clean`. This data frame will be used for plot2.
mover.clean <- na.omit(mover)

# mover1 contains 20 countries with biggest score movement. 
mover1 <- head(mover.clean, 20)

# Use pivot_longer to move score in 2011 and 2021 into rows
mover1 <- pivot_longer(data= mover1,cols= c( "year.2011", "year.2021"))
colnames(mover1) <- c("Country", "Continent", "gap.normal","gap", "year", "Score")

# Create a dataframe `mover.2011` and `mover.2021`
mover.2011 <- mover1[mover1$year == "year.2011",]
mover.2021 <- mover1[mover1$year == "year.2021",]

Create a dumbbell plot to show the different score in 2011 and 2021 (10 year difference)

# Create ggplot for data frame mover
ggplot(mover1)+
  
# Create the line of mean global in 2021. 
   geom_vline(xintercept = mean.sd$Score[2], linetype = "solid", size = 0.8, alpha = .8, color = "#009688")+
   coord_cartesian(clip = "off")+
   geom_text(x = mean.sd$Score[2]+0.05, y = 21.5, label = "Mean (2021) : 5.53", size = 4, color = "#009688")+
  
# Create the point and segment for each countries.
 geom_segment(data = mover.2021,
              aes(x = Score, y = reorder(Country, gap.normal),
              yend = mover.2011$Country, xend = mover.2011$Score), #use the $ operator to fetch data from our "data 2011"
              color = "#e3e2e1",
              size = 5,
              alpha = .5) +
  geom_point(aes(x = Score, y = Country, color = year), size = 5, show.legend = TRUE)+
  scale_color_manual(values = c("#C3B236",'#224709'),
                    labels=c("Score 2011", "Score 2021")) +

# Styling the plot with text. 
  geom_text(data=mover.2021, aes(x=Score , y=Country, label=round(Score,2)),size=4, hjust=1.5) +
  geom_text(data=mover.2011, aes(x=Score , y=Country, label=round(Score,2)),size=4, hjust=1.5, color='grey') +

# Create the gap values box.
  geom_rect(data=mover.2021, aes(xmin=7.7, xmax=8.3, ymin=0, ymax=8.5), fill="#BABD8D", colour = NA) +
  geom_rect(data=mover.2021, aes(xmin=7.7, xmax=8.3, ymin=8.5, ymax=20.5), fill='#90A955', colour = NA)+
  geom_text(data = mover.2021,aes(label = gap.normal, x = 8, y = Country), color = "white", size = 4.5)+
  geom_text(x = 8, y = 21, label = "GAP", size = 5, color = "Black")+
  
# Complete the plot with theme and labels. 
  theme_ipsum(grid="") +
  theme(plot.title = element_text(size=22),
        plot.subtitle = element_text(size = 15),
        plot.caption = element_text(size = 15),
        axis.title.x = element_text(size= 17, color = '#555955'),
        axis.text.y = element_text(size = 18, color = 'black'),
        axis.text.x = element_blank(),
        legend.text = element_text(size= 16, color = '#555955'),
        legend.title = element_blank(),
        legend.position = 'bottom') +
  labs(x= "", y=NULL, 
       title="The Biggest Life Satisfaction Score\nMovement from 2011 to 2021",
       subtitle = "Total countries in dataset is 149. However, the plot only shows\n20 countries which has the biggest score movement",
       caption  = 'Source: The World Happiness Report',
       
       ) 

Insight:

• Zimbabwe and Bahrain have the biggest positive and negative score movements, respectively.
• 12 countries have a positive score movement and 8 countries have a negative score movement.
• It’s interesting to note that ten countries’ scores have increased from the period 2011 to 2021 to above the mean global score (intersection to the mean line score). On other hand, there are only two countries whose scores have moved below the global mean. Those countries are Venezuela and Jordan.

Question 3 (b)

For data requirement, we can use data frame mover.clean and grouped by Continent.

# Create data frame `mover.con` group by continent.  
mover.con <- aggregate.data.frame(
  x=list(
    year.2011= mover.clean$year.2011, 
    year.2021= mover.clean$year.2021), 
  by = list(mover.clean$Continent), 
  FUN= mean)

# Create a new column `gap` and `gap.normal` 
mover.con$gap <- abs(mover.con$year.2021 - mover.con$year.2011)
mover.con$gap.normal <- mover.con$year.2021 - mover.con$year.2011

# Pivot_longer to move columns year.2011 and year.2021 into rows. 
mover.con <- pivot_longer(data= mover.con,cols= c( "year.2011", "year.2021"))
colnames(mover.con) <- c("Continent", "gap.normal","gap", "year", "Score")

# Create two dataframe containing data in 2011 and data in 2021.
mover.con.2011 <- mover.con[mover.con$year == "year.2011",]
mover.con.2021 <- mover.con[mover.con$year == "year.2021",]

Create a dumbbell plot

# Create ggplot for data frame mover
ggplot(mover.con)+
  
#Create the line of mean global in 2021. 
   geom_vline(xintercept = mean.sd$Score[2], linetype = "solid", size = 0.5, alpha = .6, color = "#009688")+
   coord_cartesian(clip = "off")+
   geom_text(x = mean.sd$Score[2], y = 0, label = "Mean (2021) : 5.53", size = 4, color = "#009688")+
  
# Create the point and segment for each continent.
 geom_segment(data = mover.con.2021,
              aes(x = Score, y = reorder(Continent, gap),
              yend = mover.con.2011$Continent, xend = mover.con.2011$Score), 
              color = "#e3e2e1",
              size = 5,
              alpha = 1) +
  geom_point(aes(x = Score, y = Continent, color = year), size = 5, show.legend = TRUE)+
  scale_color_manual(values = c("#C3B236",'#224709'),
                     labels=c("Score 2011", "Score 2021")) +
  xlim(4,9)+
  

# Styling the plot with text. 
  geom_text(data=mover.con.2021, aes(x=Score , y=Continent, label=round(Score,2)),size=4, hjust=1.5) +
  geom_text(data=mover.con.2011, aes(x=Score , y=Continent, label=round(Score,2)),size=4, hjust=-1, color='grey') +

# Create the background for the placement of the gap values.
  geom_rect(data=mover.con.2021, aes(xmin=8, xmax=8.6, ymin=0.5, ymax=5.5), fill="#BABD8D", colour = NA) +
  geom_rect(data=mover.con.2021, aes(xmin=8, xmax=8.6, ymin=5.5, ymax=10.5), fill='#90A955', colour = NA)+
  geom_text(data = mover.con.2021,aes(label = round(gap,2), x = 8.3, y = Continent), color = "white", size = 4.5)+
  geom_text(x = 8.3, y = 11, label = "GAP", size = 5, color = "Black")+
  
# Complete the plot with theme and labels. 
  theme_ipsum(grid="") +
  theme(plot.title = element_text(size=21),
        plot.subtitle = element_text(size = 13),
        plot.caption = element_text(size = 12),
        axis.title.x = element_text(size= 18, color = '#555955'),
        axis.text.y = element_text(size = 17, color = 'black'),
        axis.text.x = element_blank(),
        legend.text = element_text(size= 12, color = '#555955'),
        legend.title = element_blank(),
        legend.position = "bottom") +
  labs(x= "", y=NULL, title="Life Satisfaction Score Movement\nfrom 2011 to 2021 Based on Continent",
       caption = 'Source :The World Happiness Report') 

Insight:

• The average global score movement for almost 10 years is approximately (+) 0.063. It implies that more countries are ‘happier’ in the present time compared to the past.
• As it is shown in the graph, Central and Eastern Europe has significantly improved their life satisfaction score compare to other continents. Their score moved from below to above the mean value over 10 years period. On the other side, the Middle East and Africa have the lowest score movement.
• While four continent’s scores are located far from the mean values, which are scores from Sub-Saharan Africa, South Asia, Western Europe North America and ANZ, the other continent’s scores remain close with the mean value over ten years.