coding_exercise.qmd

---
title: "R Coding Exercise"
output:
  html_document:
    toc: FALSE
---

## Setting Up the Exercise

```{r}
library(tidyverse)
library(dslabs)

#look at help file for gapminder data
#help(gapminder)

#get an overview of data structure
str(gapminder)

#get a summary of data
summary(gapminder)

#determine the type of object gapminder is
class(gapminder)

#get better view of the data
glimpse(gapminder)
```

## Examining Africa Data

```{r}
africadata<- gapminder %>%
  filter(continent == "Africa")

str(africadata)
summary(africadata)

```

### Infant Mortality and Life Expectancy

``` {r}
mort_expect <- africadata %>%
  select(infant_mortality, life_expectancy, region) #kept region for sake of visualization

#examine data 
str(mort_expect)
summary(mort_expect)

#Plot
ggplot()+
  geom_point(aes(x=infant_mortality, y=life_expectancy, color=region), data=mort_expect)+ #kept region to more easily see trends since there's a lot of countries.
  xlab("Infant Mortality Rate")+ ylab("Life Expectancy (Yrs)")+
  theme_bw()
```

### Population and Life Expectancy

``` {r}
pop_expect <- africadata %>%
  select(population, life_expectancy, region) #kept region for same reason as prior plot

#examine data 
str(pop_expect)
summary(pop_expect)

#Plot
ggplot()+
  geom_point(aes(x=log(population), y=life_expectancy, color=region), data=pop_expect)+
  theme_bw()+ xlab("Log Population")+ ylab("Life Expectancy (Yrs)")

```

## Select Years
**Which years have the data missing for infant mortality?**

```{r}
mort_expect_years <- africadata %>%
  select(infant_mortality, life_expectancy, region, year)%>% #include years
  filter(is.na(infant_mortality)) #look for missing values

head(mort_expect_years) #print
```

### Select Data for 2000

Filter all Africa data for 2000

```{r}
africadata2000<-africadata%>%
  filter(year == 2000)

#Double check code 
str(africadata2000)
summary(africadata2000)
```

# Visualize the data for 2000
```{r}
#Plots
##Infant Mortality
ggplot()+
    geom_smooth(aes(x=infant_mortality, y=life_expectancy), color = "gray80", alpha = 0.1, data=africadata2000)+
  geom_point(aes(x=infant_mortality, y=life_expectancy, color=region), data=africadata2000)+
  labs(title = "Infant Mortality vs Life Expectancy", subtitle = "Year 2000") + 
  xlab("Infant Mortality")+ ylab("Life Expectancy (Yrs)") + 
  theme_bw() 

##Population
ggplot()+
    geom_smooth(aes(x=log(population), y=life_expectancy), color = "gray80", alpha = 0.1, data=africadata2000)+
  geom_point(aes(x=log(population), y=life_expectancy, color=region), data=africadata2000)+
  labs(title="Population vs. Life Expectancy", subtitle="Year 2000") + 
  xlab("Log Population")+ ylab("Life Expectancy (Yrs)") +
  theme_bw()

```

## Quantify the Data

**Fitting infant mortality as a predictor of life expectancy**

```{r}
fit1<- lm(life_expectancy ~ infant_mortality, data=africadata2000)
summary(fit1)
```

We have evidence to support that infant mortality is a predictor of life expectancy. Based on the regression model, we can predict that for every unit increase in infant mortality, the average life expectancy decreases by 0.19 years (p=2.83e-8). The average predicted life expectancy with an infant mortality rate of 0 is 71.3 years.

**Fitting population as a predictor of life expectancy**

```{r}
fit2<- lm(population ~ infant_mortality, data=africadata2000)
summary(fit2)
```

Similar to what we saw in the plots, we do not have evidence to say that population is associated with infant mortality (p=0.66). 

**This section added by Annabella Hines**

First, I wanted to look at the gapminder dataset as a whole specifically in the year 2000. I decided to create a boxplot to see the distributions of life expectancy for each continent.
```{r}
##create an object of the gapminder data for the year 2000
continent<- gapminder %>% filter(year==2000)
#A boxplot of the continent data viewing life expectancy by continent
ggplot(data=continent, aes(x=continent, y=life_expectancy, color=continent))+geom_boxplot()+xlab("Continent")+ylab("Life Expectancy")
```
The life expectancy distributions of each continent look fairly comparable except for Africa which has a lower overall distribution, lower median, and more outliers.

Next, I compared fertility to infant mortality grouped by continent for the year 2000 to see if there were any noticeable trends.

```{r}
##created a scatterplot of fertility and infant mortality color coded by continent
ggplot(data=continent, aes(x=fertility, y=infant_mortality, color=continent))+geom_point()+ylab("Infant Mortality")+ xlab("Fertility")
```
There seems to be a positive correlation between fertility and infant mortality, with Europe having low values in each and Africa having the highest. 

In the next section I wanted to explore how the life expectancy changed across the years for the different regions in Africa.
```{r}
#create an object out of africadata with year, life expectancy, and region
africaregions<- africadata %>% select(year, life_expectancy, region)
#create plot showing year vs. life expectancy color coded by region
ggplot(data=africaregions, aes(x=year, y=life_expectancy, color=region))+geom_smooth()+ylab("Life Expectancy")+xlab("Year")
```
The plot shows a relatively positive correlation between year and life expectancy, so I wanted to run a fit to verify this observation.
```{r}
#Fit life expectanct against year for the africadata
fit3<-lm(life_expectancy~year, data=africadata)
summary(fit3)
```
According to the above data, year and life expectancy for the African countries are positively correlated at the 0.05 significance level.
```{r}
#Load broom and present lm output in a table
library(broom)
map_df(list(fit3), tidy)
```