library(readr)
library(dplyr)
Attaching package: 'dplyr'The following objects are masked from 'package:stats':
filter, lagThe following objects are masked from 'package:base':
intersect, setdiff, setequal, unionlibrary(ggplot2)
library(ggcorrplot)
library(caret)Loading required package: latticechild_mortality <- read_csv("child-mortality.csv") %>%
rename(Country = Entity, Year = Year, Child_Mortality_Rate = `Child mortality rate`)Rows: 16655 Columns: 4
── Column specification ────────────────────────────────────────────────────────
Delimiter: ","
chr (2): Entity, Code
dbl (2): Year, Child mortality rate
ℹ Use `spec()` to retrieve the full column specification for this data.
ℹ Specify the column types or set `show_col_types = FALSE` to quiet this message.extreme_poverty <- read_csv("share-of-population-living-in-extreme-poverty .csv") %>%
rename(Extreme_Poverty_Share = `Share below $2.15 a day`)Rows: 2705 Columns: 3
── Column specification ────────────────────────────────────────────────────────
Delimiter: ","
chr (1): Country
dbl (2): Year, Share below $2.15 a day
ℹ Use `spec()` to retrieve the full column specification for this data.
ℹ Specify the column types or set `show_col_types = FALSE` to quiet this message.mean_income <- read_csv("mean-income-or-consumption-per-day.csv") %>%
rename(Mean_Income = `Mean income or consumption per day`)Rows: 2705 Columns: 3
── Column specification ────────────────────────────────────────────────────────
Delimiter: ","
chr (1): Country
dbl (2): Year, Mean income or consumption per day
ℹ Use `spec()` to retrieve the full column specification for this data.
ℹ Specify the column types or set `show_col_types = FALSE` to quiet this message.poorest_10 <- read_csv("threshold-income-or-consumption-per-day-marking-the-poorest-decile.csv") %>%
rename(Poorest_10_Threshold = `Threshold income or consumption per day marking the poorest decile`)Rows: 2705 Columns: 3
── Column specification ────────────────────────────────────────────────────────
Delimiter: ","
chr (1): Country
dbl (2): Year, Threshold income or consumption per day marking the poorest d...
ℹ Use `spec()` to retrieve the full column specification for this data.
ℹ Specify the column types or set `show_col_types = FALSE` to quiet this message.richest_10 <- read_csv("threshold-income-or-consumption-per-day-marking-the-richest-decile.csv") %>%
rename(Richest_10_Threshold = `Threshold income or consumption per day marking the richest decile`)Rows: 2705 Columns: 3
── Column specification ────────────────────────────────────────────────────────
Delimiter: ","
chr (1): Country
dbl (2): Year, Threshold income or consumption per day marking the richest d...
ℹ Use `spec()` to retrieve the full column specification for this data.
ℹ Specify the column types or set `show_col_types = FALSE` to quiet this message.fertility_rate <- read_csv("fertility-rate-children-per-woman.csv") %>%
rename(Country = Entity, Fertility_Rate = `Fertility rate - Sex: all - Age: all - Variant: estimates`)Rows: 18722 Columns: 3
── Column specification ────────────────────────────────────────────────────────
Delimiter: ","
chr (1): Entity
dbl (2): Year, Fertility rate - Sex: all - Age: all - Variant: estimates
ℹ Use `spec()` to retrieve the full column specification for this data.
ℹ Specify the column types or set `show_col_types = FALSE` to quiet this message.median_age <- read_csv("median-age.csv") %>%
rename(Country = Entity, Median_Age = `Median age - Sex: all - Age: all - Variant: estimates`)Rows: 18722 Columns: 3
── Column specification ────────────────────────────────────────────────────────
Delimiter: ","
chr (1): Entity
dbl (2): Year, Median age - Sex: all - Age: all - Variant: estimates
ℹ Use `spec()` to retrieve the full column specification for this data.
ℹ Specify the column types or set `show_col_types = FALSE` to quiet this message.dependency_ratio <- read_csv("total-dependency-ratio.csv") %>%
rename(Country = Entity, Dependency_Ratio = `Total dependency ratio - Sex: all - Variant: estimates`)Rows: 18944 Columns: 3
── Column specification ────────────────────────────────────────────────────────
Delimiter: ","
chr (1): Entity
dbl (2): Year, Total dependency ratio - Sex: all - Variant: estimates
ℹ Use `spec()` to retrieve the full column specification for this data.
ℹ Specify the column types or set `show_col_types = FALSE` to quiet this message.final_data <- child_mortality %>%
full_join(extreme_poverty, by = c("Country", "Year")) %>%
full_join(mean_income, by = c("Country", "Year")) %>%
full_join(poorest_10, by = c("Country", "Year")) %>%
full_join(richest_10, by = c("Country", "Year")) %>%
full_join(fertility_rate, by = c("Country", "Year")) %>%
full_join(median_age, by = c("Country", "Year")) %>%
full_join(dependency_ratio, by = c("Country", "Year"))
poverty_mortality_data <- final_data %>%
filter(Year >= 2000) %>%
filter(
!is.na(Child_Mortality_Rate),
!is.na(Extreme_Poverty_Share),
!is.na(Mean_Income),
!is.na(Poorest_10_Threshold),
!is.na(Richest_10_Threshold)
)poverty_mortality_data <- poverty_mortality_data %>%
mutate(
Region = case_when(
Country %in% c("Afghanistan", "Bangladesh", "India", "Nepal", "Pakistan", "Sri Lanka") ~ "South Asia",
Country %in% c("Algeria", "Angola", "Benin", "Botswana", "Burkina Faso", "Burundi",
"Cameroon", "Cape Verde", "Central African Republic", "Chad", "Comoros",
"Democratic Republic of Congo", "Djibouti", "Egypt", "Equatorial Guinea",
"Eritrea", "Eswatini", "Ethiopia", "Gabon", "Gambia", "Ghana", "Guinea",
"Guinea-Bissau", "Kenya", "Lesotho", "Liberia", "Libya", "Madagascar",
"Malawi", "Mali", "Mauritania", "Mauritius", "Morocco", "Mozambique",
"Namibia", "Niger", "Nigeria", "Rwanda", "Sao Tome and Principe",
"Senegal", "Seychelles", "Sierra Leone", "Somalia", "South Africa",
"South Sudan", "Sudan", "Tanzania", "Togo", "Tunisia", "Uganda",
"Zambia", "Zimbabwe") ~ "Sub-Saharan Africa",
Country %in% c("Brunei", "Cambodia", "East Timor", "Indonesia", "Laos", "Malaysia",
"Myanmar", "Philippines", "Singapore", "Thailand", "Vietnam") ~ "Southeast Asia",
Country %in% c("Argentina", "Belize", "Bolivia", "Brazil", "Chile", "Colombia",
"Costa Rica", "Cuba", "Dominican Republic", "Ecuador", "El Salvador",
"Guatemala", "Honduras", "Mexico", "Nicaragua", "Panama", "Paraguay",
"Peru", "Uruguay", "Venezuela") ~ "Latin America",
Country %in% c("China", "Japan", "Mongolia", "North Korea", "South Korea", "Taiwan") ~ "East Asia",
Country %in% c("Bahrain", "Iran", "Iraq", "Israel", "Jordan", "Kuwait", "Lebanon",
"Oman", "Palestine", "Qatar", "Saudi Arabia", "Syria", "United Arab Emirates",
"Yemen") ~ "Middle East",
Country %in% c("Albania", "Austria", "Belarus", "Belgium", "Bosnia and Herzegovina", "Bulgaria",
"Croatia", "Czechia", "Denmark", "Estonia", "Finland", "France", "Germany",
"Greece", "Hungary", "Ireland", "Italy", "Kosovo", "Latvia", "Lithuania",
"Moldova", "Netherlands", "North Macedonia", "Norway", "Poland", "Portugal",
"Romania", "Russia", "Serbia", "Slovakia", "Slovenia", "Spain", "Sweden",
"Switzerland", "Ukraine", "United Kingdom") ~ "Europe",
Country %in% c("Kazakhstan", "Kyrgyzstan", "Tajikistan", "Turkmenistan", "Uzbekistan") ~ "Central Asia",
Country %in% c("Australia", "Fiji", "Kiribati", "Marshall Islands", "Micronesia (country)",
"Nauru", "New Zealand", "Palau", "Papua New Guinea", "Samoa", "Solomon Islands",
"Tonga", "Tuvalu", "Vanuatu") ~ "Oceania",
TRUE ~ NA_character_
)
)regional_mortality <- poverty_mortality_data %>%
filter(Year >= 2020 & !is.na(Region)) %>%
group_by(Region) %>%
summarise(Average_Mortality = mean(Child_Mortality_Rate, na.rm = TRUE)) %>%
arrange(desc(Average_Mortality))region_colors <- c(
"Sub-Saharan Africa" = "#D7263D", # Red
"South Asia" = "#66A61E", # Green
"Southeast Asia" = "#1C9EDB", # Blue
"Latin America" = "#FF8C42", # Orange
"East Asia" = "#7E57C2", # Purple
"Middle East" = "#FFB74D", # Light Orange
"Europe" = "#42A5F5", # Light Blue
"Central Asia" = "#009688", # Teal
"Oceania" = "#AB47BC" # Violet
)ggplot(regional_mortality, aes(x = reorder(Region, Average_Mortality), y = Average_Mortality, fill = Region)) +
geom_col() +
geom_text(aes(label = round(Average_Mortality, 2)), hjust = -0.1, size = 6, color = "black") +
scale_fill_manual(values = region_colors) +
coord_flip() +
labs(
title = "Average Under-5 Child Mortality Rate by Region (2020–2023)",
x = "Region",
y = "Mortality Rate (deaths per 1,000 live births)",
caption = "Source: Our World in Data, UN Population Division"
) +
theme_minimal(base_size = 13) +
theme(legend.position = "none")
regional_share <- regional_mortality %>%
mutate(
Share = Average_Mortality / sum(Average_Mortality),
Label = paste0(Region, "\n", round(Share * 100, 1), "%")
)
ggplot(regional_share, aes(x = "", y = Share, fill = Region)) +
geom_col(width = 1, color = "white") +
coord_polar(theta = "y") +
geom_text(aes(label = paste0(round(Share * 100, 1), "%")),
position = position_stack(vjust = 0.5), size = 2, color = "white") +
scale_fill_manual(values = region_colors) +
labs(
title = "Child Mortality Rate Share by Region (2020–2023)",
caption = "Source: Our World in Data, UN Population Division"
) +
theme_void(base_size = 13) +
theme(legend.title = element_blank())
scatter_data1 <- poverty_mortality_data %>%
filter(Year >= 2020, !is.na(Mean_Income), !is.na(Child_Mortality_Rate), !is.na(Region))
ggplot(scatter_data1, aes(x = Mean_Income, y = Child_Mortality_Rate, color = Region)) +
geom_point(size = 2.8, alpha = 0.85) +
geom_smooth(method = "lm", se = FALSE, color = "black", linetype = "dashed") +
scale_color_manual(values = region_colors) +
labs(
title = "Higher Income is Linked to Lower Child Mortality",
subtitle = "Sub-Saharan Africa stands out with low income and high mortality rates",
x = "Mean Daily Income (USD)",
y = "Under-5 Mortality Rate (per 1,000 live births)",
caption = "Each dot represents a country | Source: Our World in Data"
) +
theme_minimal(base_size = 13) +
theme(legend.title = element_blank())`geom_smooth()` using formula = 'y ~ x'
scatter_data2 <- poverty_mortality_data %>%
filter(Year >= 2020, !is.na(Fertility_Rate), !is.na(Child_Mortality_Rate), !is.na(Region))
ggplot(scatter_data2, aes(x = Fertility_Rate, y = Child_Mortality_Rate, color = Region)) +
geom_point(size = 2.8, alpha = 0.85) +
geom_smooth(method = "lm", se = FALSE, color = "black", linetype = "dashed") +
scale_color_manual(values = region_colors) +
labs(
title = "Higher Fertility Is Associated with Higher Child Mortality",
subtitle = "Sub-Saharan Africa again shows clustering of high fertility and mortality rates",
x = "Fertility Rate (Children per Woman)",
y = "Under-5 Mortality Rate (per 1,000 live births)",
caption = "Each dot represents a country | Source: Our World in Data"
) +
theme_minimal(base_size = 13) +
theme(legend.title = element_blank())`geom_smooth()` using formula = 'y ~ x'
# Select Y, X, and M variables from the final cleaned dataset
corr_data <- poverty_mortality_data %>%
select(
Child_Mortality_Rate, # Y variable
Extreme_Poverty_Share, Mean_Income, # X variables
Poorest_10_Threshold, Richest_10_Threshold,
Fertility_Rate, # X variable
Median_Age, Dependency_Ratio # M variables
)
# Compute correlation matrix (using complete observations)
cor_matrix <- cor(corr_data, use = "complete.obs")
# Plot the correlation heatmap
ggcorrplot(
cor_matrix,
method = "square",
type = "lower",
lab = TRUE,
lab_size = 3,
colors = c("#D7263D", "white", "#1C9EDB"), # Red to white to blue scale
title = "Correlation Heatmap: Predictors & Under-5 Mortality",
ggtheme = ggplot2::theme_minimal()
)
# Filter Relevant Variables
model_data <- poverty_mortality_data %>%
select(
Child_Mortality_Rate,
Extreme_Poverty_Share,
Mean_Income,
Poorest_10_Threshold,
Richest_10_Threshold,
Fertility_Rate,
Median_Age,
Dependency_Ratio
) %>%
na.omit() # Remove missing rows
# Train-Test Split (70/30)
set.seed(123) # for reproducibility
train_index <- createDataPartition(model_data$Child_Mortality_Rate, p = 0.7, list = FALSE)
train_data <- model_data[train_index, ]
test_data <- model_data[-train_index, ]
# Fit Base Model (Multiple Linear Regression)
base_model <- lm(Child_Mortality_Rate ~ ., data = train_data)
# Predict on Test Set
predictions <- predict(base_model, newdata = test_data)
# Calculate R-squared on Test Set
r_squared <- cor(predictions, test_data$Child_Mortality_Rate)^2
cat("R-squared (Test Set):", round(r_squared, 3), "\n")R-squared (Test Set): 0.835 # Optional: View model summary
summary(base_model)
Call:
lm(formula = Child_Mortality_Rate ~ ., data = train_data)
Residuals:
Min 1Q Median 3Q Max
-5.1262 -0.4587 -0.0021 0.3795 15.8635
Coefficients:
Estimate Std. Error t value Pr(>|t|)
(Intercept) -0.819027 0.445432 -1.839 0.066213 .
Extreme_Poverty_Share 0.084323 0.004145 20.342 < 2e-16 ***
Mean_Income -0.201833 0.054192 -3.724 0.000205 ***
Poorest_10_Threshold 0.131564 0.037889 3.472 0.000535 ***
Richest_10_Threshold 0.078735 0.023254 3.386 0.000734 ***
Fertility_Rate 1.780150 0.110252 16.146 < 2e-16 ***
Median_Age 0.016602 0.009912 1.675 0.094208 .
Dependency_Ratio -0.031734 0.007559 -4.198 2.9e-05 ***
---
Signif. codes: 0 '***' 0.001 '**' 0.01 '*' 0.05 '.' 0.1 ' ' 1
Residual standard error: 1.285 on 1152 degrees of freedom
Multiple R-squared: 0.8135, Adjusted R-squared: 0.8124
F-statistic: 718.1 on 7 and 1152 DF, p-value: < 2.2e-16# Calculate RMSE
rmse <- sqrt(mean((predictions - test_data$Child_Mortality_Rate)^2))
# Model coefficients summary
model_coef <- summary(base_model)$coefficientsThis analysis demonstrates how income, fertility, and poverty related indicators strongly predict under-5 mortality across global regions particularly Sub-Saharan Africa and South Asia.