Lab1

source("http://www.openintro.org/stat/data/cdc.R")
# Lab 1 OpenIntro Stats
# Intro to Data
names(cdc)
length(cdc)
str(cdc)
head(cdc)
tail(cdc)
# Ex. 1
summary(cdc$weight)
190.0 - 140.0
mean(cdc$weight)
var(cdc$weight)
median(cdc$weight)
table(cdc$smoke100)
table(cdc$smoke100)/20000
barplot(table(cdc$smoke100))
# Ex. 2
summary(cdc$height)
70-64 # interquartile range of height
summary(cdc$age)
57-31 # interquatile range of age
table(cdc$gender)
table(cdc$exerany)
table(cdc$gender)/20000 # rel.freq dist for gender
table(cdc$exerany)/20000
table(cdc$genhlth)
table(cdc$genhlth)/20000
mosaicplot (table(cdc$gender,cdc$smoke100))
dim(cdc)
cdc[567,6]
names(cdc)
mdata <- subset(cdc, cdc$gender == "m")
head(mdata)
Under23Smokedata <- subset(cdc, cdc$smoke100 == "1" & cdc$age < 23)
head(Under23Smokedata)
boxplot(cdc$height)
boxplot(cdc$height ~ cdc$gender)
bmi <- (cdc $ weight / cdc $ height ^2) * 703
boxplot (bmi ~ cdc $ genhlth )
hist(cdc$age)
hist(bmi)
hist(bmi, breaks = 50)
# Ex: On Your Own
# 1. Scatterplot of weight vs. desired weight
names(cdc)
plot(cdc$weight ~ cdc$wtdesire)
library("car")
plot(cdc$weight, cdc$wtdesire)
scatterplot(cdc$weight ~ cdc$wtdesire)
pairs(~cdc$weight+cdc$wtdesire, data=cdc, main="CDC weight vs desired weight")
# In general, desired weight of the respondents is less than actual weight.
# 2. New variable: wdiff
wdiff <- cdc$wtdesire - cdc$weight
str(wdiff)
# 3. wdiff data is integer. if = 0 then person is ok with own weight. + or - relates to person's own eval of weight
plot(wdiff)
plot(cdc$weight, cdc$wtdesire)
hist(wdiff, breaks = 25)
fivenum(wdiff)
summary(wdiff) # 4. In general, people want to lose 10 to 15 pounds. Centered at 10lbs to lose, the wdiff is normally distributed with a range of 21lbs.
# 6.