A.
Reading Bar Charts
This chart shows the number of Starbucks cafes in three states — New Jersey, New York, and
Pennsylvania. What state has about 400 cafes? What state has about 125 cafes? What state has
the most cafes? What state has the fewest cafes?
Some questions are easier to answer by organizing the data in different ways.
What pair of states have the most cafes combined? Which state has a number of cafes most
similar to PA?
Bar charts are best for categorical data — data that fall into distinct groups, like occupation, type
of vehicle, or US state.
Practice
Which drink has the highest sales?
Which drink did the cafe sell about 250 of?
Which 2 drinks sold about 225 combined?
B. Analyzing Bar Charts
This chart shows the number of Starbucks cafes in five US states.
What state has the largest number of cafes?
About how many more cafes are in MA compared to NH?
a/ 100 b/ 250 c/ 400
To make other comparisons, let's remove MA from the chart.
About how many more cafes are in NH compared to RI?
a/ 10 b/ 20 c/ 30
Excluding MA, we were able to see the small differences between the remaining
states.
Removing outliers can reveal important details about the rest of the data on a bar
chart.
Select the outlier.
Removing CA, seven states remain on the chart.
Which states have fewer than 100 cafes?
Which state has the most similar number of cafes to OR?
Data visualization is equal parts art and science. You'll have to weigh the pros and
cons of excluding outliers from your visualizations.
Practice
Select the outlier.
Select which drink had the highest sales out of mochas, espressos, and lattes.
About how many more espressos were sold compared to mochas?
a/ 40 b/ 60 c/ 80
C. Pie Charts
A common way to report a category's size is to compare it to the total as a
percentage — let's visualize it.
This bar chart compares the number of Starbucks cafes in states on the Great
Plains.
About how many total cafes are there in these 5 states?
a/ 1000 b/ 1500 c/ 2000
Knowing there are about 1000 cafes total, what state has about half of the cafes in
this region?
Knowing there are about 1000 cafes total, what three states combined have about
30% of the cafes in this region?
While the volume of cafes is easy to see in a bar chart, the percent of the whole is
not. Percents are more easily read from a pie chart
What pair of states have about 70% of the cafes?
What state has 20% of the total number of cafes in this region?
Pie charts show how each category contributes to the total.
Practice
What percent of the total sales do cappuccinos have?
What percent of the total sales do mochas (M) have?
a/ 10% b/ 25% c/ 50%
What percent of the total sales do espressos and lattes have together?
a/ 25% b/ 50% c/ 70%
D. Histograms
We can visualize a distribution of ages — or any variable that has lots of possible
values – on a special kind of bar chart called a histogram.
This bar chart shows the ages of 800 Starbucks customers, split into two groups.
How could we improve this chart to learn more about Starbucks customers' ages?
a/ Use more age groups b/ Sample more customers
Here's the same data, but broken into more age groups.
What 15-year age range has the largest number of Starbucks customers?
The different age ranges are called bins. The width of each bin is 15, and all bins
are the same width.
What would give us a more precise estimate of the peak age?
a/ Narrower bins b/ Wider bins
Here's the same data with narrower bins.
What is the most common 5-year age range in Starbucks?
a/ 45 to 50 b/ 50 to 55 c/ 55 to 60
Data that take lots of possible values, like age, are best displayed on histograms —
bar charts with bins that have equal widths.
In our sample, there are 60 customers aged 18 to 30 — a group that's crucial for the
long-term health of a company like Starbucks.
To compare Starbucks' popularity among ages 18-22 and ages 26-30, what change
could be made to the bins?
a/ Make them narrower b/ Make them wider
Histograms that display smaller samples require wider bins.
● The bin width should be narrow enough to see important details.
● The bin width should be wide enough to smooth out an overall pattern.
Making better business decisions is one reason to visualize customer data on a
histogram.
Practice
What is the most common 50-minute range of time spent on leisure?
In how many countries do people spend between 250 and 300 minutes on leisure?
Approximately what percentage of the 33 countries spend less than 250 minutes on
leisure?
a/ 12% b/ 25% c/ 41%
E. Percentiles
A well-targeted product should appeal to a large number of people, but it shouldn't
aim to please everyone.
We'll use percentiles to choose which people to target.
We've collected data on Starbucks' customers and made a histogram of their ages:
If you were selling a new product to Starbucks customers, what age would make the
most sense to focus on?
a/ 15 to 40 b/ 40 to 65 c/ 65+
It makes sense to target age 56, which is the median (trung vị) age or 50th
percentile — 56 is older than 50% of customers.
What percentage of Starbucks customers are aged under 40?
What percentage of customers are aged between 25 and 50?
a/ 15% b/ 20% c/ 30% d/ 45%
The age range from 25 to 50 spans 25 years, but it only includes 30% of customers.
We'll do better if we focus on an age range around the median age of 56.
What percentage of customers are aged between 43 and 67?
By focusing on the center of the age distribution we can reach 50% of customers
with an age range of 67−43=24 years.
The range from the 25th to 75th percentile is known as the interquartile range.
Age isn't the only factor to consider. We want to be sure to target customers who can
afford our product.
What is the interquartile age range of Starbucks customers with an income of at least
$90,000?
Percentiles can help us find ranges where a continuous variable is concentrated.
Practice
In what percentage of countries is leisure time less than 250 minutes?
What is the median amount of minutes a country spends on leisure?