R Notebook

setwd(“~/career_da-pffs/udacity-eda/eda-course-materials/lesson6”)

str(gDat)

'data.frame':   1680 obs. of  6 variables:
 $ country  : Factor w/ 140 levels "Afghanistan",..: 1 1 1 1 1 1 1 1 1 1 ...
 $ year     : int  1952 1957 1962 1967 1972 1977 1982 1987 1992 1997 ...
 $ pop      : num  8425333 9240934 10267083 11537966 13079460 ...
 $ continent: Factor w/ 4 levels "Africa","Americas",..: 3 3 3 3 3 3 3 3 3 3 ...
 $ lifeExp  : num  28.8 30.3 32 34 36.1 ...
 $ gdpPercap: num  779 821 853 836 740 ...

#summary(gDat)
cor.test(gDat$lifeExp, gDat$gdpPercap, method = 'pearson') 

    Pearson's product-moment correlation

data:  gDat$lifeExp and gDat$gdpPercap
t = 28.799, df = 1678, p-value < 2.2e-16
alternative hypothesis: true correlation is not equal to 0
95 percent confidence interval:
 0.5422206 0.6062796
sample estimates:
      cor 
0.5751311 

Observation - a strong positive correlation.

summary(gDat$lifeExp) Min. 1st Qu. Median Mean 3rd Qu. Max. 23.60 48.08 60.34 59.26 70.75 82.60

1. A scatter plot of GDPPercap vs LifeExp

ggplot(aes(x = lifeExp, y = gdpPercap, color = continent), data = gDat) + 
  geom_point() +
  ggtitle("GdpPercap vs. LifeExp") +
  labs(x="lifeExp", y="gdpPercap") +
  scale_x_continuous(breaks = seq(20,90,4)) + 
  xlim(20, quantile(gDat$lifeExp, 0.99)) 

Scale for 'x' is already present. Adding another scale for 'x', which will replace the
existing scale.

Observation - A few outliers - a country with almost 0 gdpPercap and some with extremely high GDP Most countries with a gdp below 15000. To reduce overploting with alpha and coordinate transormation

min(gDat$gdpPercap)

ggplot(aes(x = lifeExp, y = gdpPercap, color = continent), data = gDat) + 
  geom_point(alpha = 1/2) +
  ggtitle("GdpPercap vs. LifeExp") +
  labs(x="lifeExp", y="gdpPercap") +
  scale_x_continuous(breaks = seq(20,90,4)) + 
  coord_trans(x="sqrt", y='sqrt')

Omiting the top 1% of values by applying the quartile method

ggplot(aes(x = lifeExp, y = gdpPercap, color = continent), data = gDat) + 
  geom_point(alpha = 1/2) +
  ggtitle("GdpPercap vs. LifeExp") +
  labs(x="lifeExp", y="gdpPercap") +
  scale_x_continuous(breaks = seq(20,90,4)) + 
  coord_trans(x="sqrt", y='sqrt') + 
   xlim(20, quantile(gDat$lifeExp, 0.99)) +
  ylim(0, quantile(gDat$gdpPercap, 0.99)) +
  geom_smooth(method="lm", color="red")

Scale for 'x' is already present. Adding another scale for 'x', which will replace the
existing scale.

Geom-smooth omits most of the values therefore does not seem the proper choice for this data

Making the plot more readable by creating and using conditoning summaries.
Error below - I do not understand it??

gDat_by_country<- gDat %>% 
                group_by(country) %>%
                summarise(lifeExp_mean=mean(lifeExp),
                lifeExp_median=median(lifeExp),
                 n=n()) %>%
                arrange(country)

Error: This function should not be called directly
Call `rlang::last_error()` to see a backtrace

ggplot(aes(x = continent, y = lifeExp),
       data = subset(gDat, !is.na(continent))) + 
  geom_boxplot() +
  stat_summary(geom = 'point', shape = 4, fun.y = mean)

ggplot(aes(x = continent, y = gdpPercap, color = continent),
       data = subset(gDat, !is.na(continent))) + 
  geom_boxplot() +
  stat_summary(geom = 'point', shape = 4, fun.y = median) +
  scale_y_continuous(limits = c(0, 30000),breaks=seq(0, 30000, 1000))

NA

Write code to create a new data frame - why??

year_lifeExp_group <- group_by(gDat, year, lifeExp) %>% 
                filter(!is.na(year)) %>% 
                summarise(mean_gdpPercap = mean(gdpPercap),
                          median_gdpPercap = median(gdpPercap),
                          n = n())

Error: This function should not be called directly
Call `rlang::last_error()` to see a backtrace

``` why?