Always start by looking at the distributions of variables: center, shape, spread, and weird stuff.
- Package:
okcupiddata - Data frame:
profiles
Find out about the data frame profiles. How many observations and variables does the profiles data frame have?
str(profiles)
- 59946 observations, 22 variables
Examine the distribution of age.
gf_histogram(~ age, data=profiles, fill="orange")
Why do you think this distribution has the shape it does?
- unimodal
- not normal
- skewed right
Run a favstats() on age.
favstats(~age, data=profiles)
- What do you notice? Were there any clues in the histogram?
- The max age is 110. The minimum is 18. The median is 30 (middle score in whole distribution).
- Suggest some ways to figure out what's going on, and how to fix it.
profiles <- filter(profiles, age<100)
favstats(~age, data=profiles)
gf_histogram(~age, data=profiles, fill="purple")
favstats(~ age, data=profiles)
- Plot the five number summary for age on a number line - to scale
- min - Q1 - median - Q3 - max
- Why are the quartiles not equally spaced?
- the intervals contain the same number of values
- Sketch the boxplot of this distribution
gf_bar(~ bar, data=profiles, fill="orange")
- looks just like a histogram, but produced using a different method
- shape distribution doesn't make sense for a bar graph
- In the profiles data frame, does Sex explain Height?
- fewer females than males
- Working Definition: Knowing someone's value on the explanatory variable helps you make a better guess about their value on the outcome variable.