Chapter 2: Exploratory Data Analysis

what are the 2 ways to visualize categorical data?

bar plot

pie chart

what data can be represented in bar plot?

any. can have more than 1 choice.

what data can be represented in pie chart?

data that only has one choice.

what is on the horizontal axis of bar plots?

categories

what is on the vertical axis of bar plots?

counts

pie chart

convert data into percentages

what are the 2 ways to visualize quantitative data?

histogram & box-plot

how is a histogram organized?

x = variable

y = frequency

mean

average

median

middle

mode

most common

are the mean, median, & mode always enough to tell you about something?

no, you often need standard deviation

standard deviation (SD)

how each observation is deviated from the mean

small: close to mean

large: far from mean

what are the 3 diff. shapes of a histogram?

symmetric, right-skewed/tailed, left-skewed/tailed

symmetric histogram distribution

bell shape, mirror image

mean = median

right-skewed/tailed distribution

long tail on right

mean > median

left-skewed/tailed distribution

long tail on left

mean < median

in R, when do you use the table function?

only for categorical data

what is the equal sign in R?

<-

5 number summary

min, Q1, med, Q3, max

Q1

first quartile

median of first half

Q3

third quartile

median of second half

IQR (inter quartile ranfge)

Q3 -Q1

spread/variable measurements

IQR & SD

mean, median, mode

typical/center measurements

typical & spread measurement for symmetric distribution

mean & SD

typical & spread measurement for skewed distribution

median & IQR

outlier

an extreme value

formulas for finding outliers

x > Q3 + 1.5(IQR)

x < Q1 - 1.5(IQR)

how can you visualize both quantitative and categorical data?

use parallel box plot

when determining the shape of a distribution, should you look at each individual category?

NO. only look at all the data,

when do you use z-score

when you have a symmetric distribution

z-score calculation

data - mean / SD

what does z-score tell you?

how many standard deviations a specific data point is away from the mean of its dataset

compared to a skewed distribution, what does the SD & IQR of a bell-shaped curve look like?

smaller, since most points are concentrated close to center