notebook.qmd

---
title: "World Happiness Report"
format: html
editor: visual
---

```{r}
library(readr)
library(countrycode)
library(dplyr)
library(GGally)
library(visdat)
library(skimr)
library(DataExplorer)
library(corrplot)
library(lattice)
library(car)        
library(lme4)       
library(tidyverse)
```

## Introduction

The World Happiness Report is a publication from the Sustainable Development Solution Network, which measures the happiness score by considering responses from a subsample of the worldwide population.

The happiness score is mainly based on people's satisfaction with their lives; the main factor that contributes to the satisfaction are six:

-   GDP (Gross Domestic Product): how much each country produce, divided by the number of people in the country;

-   Social Support: having someone to count on in case of trouble;

-   Healthy Life Expectancy: physical and mental health status;

-   Freedom to make life choices refers to human rights, including sex, ethnicity, nationality, language, and religion;

-   Generosity: measured in how much people donate to charity;

-   Perception of corruption: if people trust in the government and the benevolence of others.

A helpful tool to predict the population's happiness from those characteristics can be implemented using a Machine Learning model. For this reason, I implemented a **Linear Mixed Effect model** to predict the happiness score of the continent for the year 2022.

## Exploratory Data Analysis and Data Preprocessing

```{r}
data <- read_delim(file.choose(), delim = ";", escape_double = FALSE, trim_ws = TRUE)
head(data)
```

The original dataset contains 2199 observations of 11 variables: country name, year, life ladder, generosity, GDP, social support, life expectancy, freedom, corruption, and negative and positive effects.

As a first step, I extracted the continent from the country name to decrease the grouping variable into size 5: Africa, Asia, Europe, North America, Oceania, South America, and Antarctica.

```{r}
df <- data %>%
  mutate(continent_code = countrycode(Country name, "country.name", "continent")) %>%
  mutate(continent = case_when(
    continent_code == "AF" ~ "Africa",
    continent_code == "AS" ~ "Asia",
    continent_code == "EU" ~ "Europe",
    continent_code == "NA" ~ "North America",
    continent_code == "OC" ~ "Oceania",
    continent_code == "SA" ~ "South America",
    continent_code == "AN" ~ "Antarctica",
    TRUE ~ "Unknown"
  ))


df <- df[2:12]
df$continent_code <- as.factor(df$continent_code)
df <- na.omit(df)
head(df)
```