SOCR Data - 25,000 Records of Human Heights (in) and Weights (lbs)

SOCR_Data_25000_Human_Height_Weight.R

## Height and Weight of 18 year olds
## from Hong Kong 1993 Growth Survey data, 
## simulated by SOCR from reported summary statistics
## Heights in inches
## Weights in pounds
## Explanation \url{http://wiki.stat.ucla.edu/socr/index.php/SOCR_Data_Dinov_020108_HeightsWeights}
## Data \url{http://socr.ucla.edu/docs/resources/SOCR_Data/SOCR_Data_Dinov_020108_HeightsWeights.html}

## Libraries ####
library(rvest)    # Web scraping
library(magrittr) # Data wrangling
library(dplyr)    # Data wrangling
library(ggplot2)  # Plotting distribution
library(nortest)  # used to test for normality

## Scrape the data from SOCR's website ####
url <- "http://socr.ucla.edu/docs/resources/SOCR_Data/SOCR_Data_Dinov_020108_HeightsWeights.html"

the_data <- read_html(url) %>% 
  html_table(header = TRUE) %>% 
  extract2(1) %>% 
  setNames(make.names(names = colnames(.))) %>% 
  mutate(Index = NULL)

## Histograms ####
## for height
the_data %>%
  ggplot(aes(x = Height.Inches.)) +
  geom_histogram(aes(y = ..density..)) +
  stat_function(fun = dnorm, args = list(mean = mean(the_data$Height.Inches.), sd = sd(the_data$Height.Inches.)), color = "blue") +
  theme_minimal()

## for weight
the_data %>%
  ggplot(aes(x = Weight.Pounds.)) +
  geom_histogram(aes(y = ..density..)) +
  stat_function(fun = dnorm, args = list(mean = mean(the_data$Weight.Pounds.), sd = sd(the_data$Weight.Pounds.)), color = "blue") +
  theme_minimal()

## Summary of data ####
summary(the_data$Height.Inches.)
summary(the_data$Weight.Pounds.)
# Note results inconsistent with "age 6 to 18"

## Demonstrating the data is simulated from a normal distribution; not actual measured data ####
ad.test(the_data$Height.Inches.)
ad.test(the_data$Weight.Pounds.)
sf.test(sample(x = the_data$Height.Inches., size = 1000, replace = FALSE))
sf.test(sample(x = the_data$Weight.Pounds., size = 1000, replace = FALSE))
lillie.test(the_data$Height.Inches.)
lillie.test(the_data$Weight.Pounds.)