Lecture 8

Histograms
Announcements
● HW02 is due today
● Lab03 is posted on data science server
○ Due coming Wednesday before class
● HW03 is posted on data science server
● Due coming Friday before class
Weekly Goals
● Monday
○ Table review
○ Working with Census data
● Wednesday
○ Visualizing data
○ Line plots, scatter plots, bar charts
● Today
○ Visualizing two kinds of distributions
○ Proportions as areas
Distributions
Terminology
● Individuals: those whose attributes are recorded
● Variable: an attribute (column)
○ can be numerical or categorical
○ has different values
■ each individual has exactly one value
○ has a distribution:
■ For each different value of the variable, the
frequency of individuals that have that value
A Distribution
Each individual is in exactly one category. Percents add up to 100.

Source: Pew Research

Not a Distribution
Percents of survey respondents on “a major reason they would find it
difficult to quarantine themselves for at least 14 days”

Each respondent
can pick more than
one answer.

The bars represent

overlapping groups.

Source: Pew Research

Categorical Distributions

(Demo)
Bar Chart
To display all the values of the variable along with all their
frequencies
● Bar chart
○ One bar for each category
○ You can choose the order of the bars
○ Length of bar is the percent (or count) of individuals in
that category

(Demo)
Numerical Distributions
Grouping Numerical Values: Binning
Binning is counting the number of numerical values that lie
within ranges, called bins.
● Bins are defined by their lower bounds (inclusive)
● The upper bound is the lower bound of the next bin
188, 170, 189, 163, 183, 171, 185, 168, 173, ... The [185,190) bin

160 165 170 175 180 185 190

(Demo)
Area Principle
What Is Wrong With This Picture?

Caption: The new iPad battery is 70% bigger than the previous iPad. Source
Area Principle
Areas should be proportional to the values they represent.

For example
● If you represent 20% of a population by

● Then 40% can be represented by:

● But not by:

Drawing Histograms
Histogram
● Displays the distribution of a numerical variable
● One bar corresponding to each bin
● Uses the area principle:
○ The area of each bar is the percent of individuals in the
corresponding bin

● Later in the course, we will

approximate histograms by
smooth curves. 68%
Areas will still represent
percents. (Demo)
Density
Histogram Axes
● plt.hist uses a scale (density=True) that ensures
the area of the chart sums to 100%

● The area of each bar is a percentage of the whole

● The horizontal axis is a number line (e.g. years), and

the bins sizes don’t have to be equal to each other

● The vertical axis is a rate (e.g., percent per year)

(Demo)
How to Calculate Height
The [40, 65) bin contains 56 out of 200 movies
● “56 out of 200” is 28%
● The bin is 65 - 40 = 25 years wide
28 percent
Height of bar = --------------------------
25 years
= 1.12 percent per year
Height Measures Density
% in bin
Height = ---------------------
width of bin

● The height measures the percent of data in the bin

relative to the amount of space in the bin.
● Height measures crowdedness, or density.
● Units: percent per unit on the horizontal axis
Area Measures Percent

Area of bar = % in bin = Height x width of bin

● “How many individuals in the bin?” Use area.

● “How crowded is the bin?” Use height.

Discussion Questions
Compare the bins [10, 20) and
[20, 40).

● Which one has more movies?

Answer: [20, 40), bigger area

● Which one is more crowded?

Answer: [10, 20), taller
Bar Chart or Histogram?
To display a distribution:
Bar Chart Histogram
● Distribution of categorical ● Distribution of numerical
variable variable
● Bars have arbitrary ● Horizontal axis is numerical:
(but equal) widths and drawn to scale, no gaps,
spacings; in any order bins can be unequal
● height (or length) and ● Area of bars proportional to
area of bars proportional to the percent of individuals;
the percent of individuals height measures density
 2016 Income
Name (millions)

Discussion Questions Jennifer Lawrence

Scarlett Johansson
61.7
57.5
Angelina Jolie 40
Jennifer Aniston 24.75
What is the height of each bar in this Anne Hathaway 24

histogram? Melissa McCarthy

Bingbing Fan
24
20
Sandra Bullock 20
Cara Delevingne 15
my_bins = make_array(0, 15, 25, 85) Reese Witherspoon 15
plt.hist( , bins = my_bins) Amy Adams 15
Kristen Stewart 12
Amanda Seyfried 10.5
Tina Fey 10.5
What are the vertical axis units? Julia Roberts 10
Emma Stone 10
Natalie Portman 8.5
Margot Robbie 8
Meryl Streep 6
Mila Kunis 4.5
 2016 Income
Name (millions)

Answers Jennifer Lawrence

Scarlett Johansson
61.7
57.5
Angelina Jolie 40
Jennifer Aniston 24.75
Vertical axis units: Percent per million Anne Hathaway 24
Melissa McCarthy 24
my_bins = make_array(0,15,25,85) Bingbing Fan 20
Sandra Bullock 20
[0, 15): (45%)/(15 million) Cara Delevingne 15

= 3 % per million Reese Witherspoon 15

Amy Adams 15
[15, 25): (40%)/(10 million) Kristen Stewart 12
Amanda Seyfried 10.5
= 4 % per million Tina Fey 10.5
Julia Roberts 10
[25, 85): (15%)/(60 million) Emma Stone 10
Natalie Portman 8.5
= 0.25 % per million Margot Robbie 8
Meryl Streep 6
Mila Kunis 4.5