Visualizing data in Spreadsheet

A. Pie charts
Illustrating proportionality

A pie chart is used to illustrate proportionality. Think of it as slicing the pie into pieces, where
each piece matches a percentage of the whole list.

In spreadsheets, this is really easy, because all we need is a list of the categories and matching
values such as sums or counts.

When the chart is selected, a design and format menu is available on the Excel ribbon at the top
of the page. The design menu gives numerous chart options and choices, such as specific
coloring or displaying percentages. You can also change the chart title.

Pivot table

Using the pivot table we created earlier with some careful selection, we want to highlight the
position categories in the top row and the totals are in the bottom row. To do this:

1. Select the categories at the top of the table.

2. While holding down the control key on Windows or the Command key on Apple
keyboards, select the bottom row with your mouse.

3. We are going to copy this highlighted data to another location.

4. Paste (paste-transpose) using the transpose features so it creates columns instead of

rows.

5. Select and choose "insert pie chart" as before.

B. Bar charts
We could use the same information as before and choose a bar or column chart instead of a pie
chart. Instead of percentages, it would just show the values with longer bars or columns
representing larger values.

Bar charts do not show percentages for each category - here it shows columns by baseball
position, with numbers on the vertical axis

Bar charts do not show percentages for each category

In the chart above, we're comparing the category values against each other and we see their
relative sizes. However, we do not have much sense of the whole league or the percentage of
each category as we did with the pie charts.

Choosing which kind of chart to use really depends on what patterns you want to highlight and
what questions you want to answer.

Bar or pie chart?

Use bar or column charts to compare category values with each other.

Use a pie chart to show the proportionality of categories.

C. Scatter or Line plots

Line Charts vs. Pie Charts

We use pie and bar charts to visualize categorical data. If we have a list of numerical data, such
as the list of stock prices over time, a line chart gives us a better picture of the data set.

Simple line charts

Using a table of data downloaded from a financial website, listing prices for AAPL stock, we can
explore line charts:

Notice that it has columns for date, open, high, low, close, and volume.

Select the date column and the close column .

Go to the insert menu and select a line chart.

Move the chart to its own sheet to see the detail better.

Choose a quick style fix in the design menu.

Change the title to AAPL Stock Price.

Verify that the horizontal axis shows the dates, and the vertical axis shows dollar values.

Observe that over the past year the stock has gone up with a little hiccup about a month ago.

Multiple columns of data

To handle more than one column of data for the same dates:

Observe lines for each on the same chart can be shown.

Select the date plus the high and low values for AAPL stock.

Since the high and low aren't all that far apart, change the range for the dollar amount on the left
to start at 100.

Select the vertical axis, right-click, and format axis.

Observe both the high and low-value lines now and see the spreads between them.

Scatterplot

To plot two different variables, closing price, and volume, for AAPL stock, choose the
scatterplot.

Observe a graph with the closing price on the horizontal axis and the volume of trade that day on
the vertical axis.

Observe that the prices seem to cluster in a couple of areas, and that they have about the same
volume generally, though there are some high volume days at the lower price.

Scatter plots are useful for displaying bivariate(opens in a new tab) numerical data. This means a
data set with two variables, such as height and weight measurements for a list of human beings.

If the data of both variables move up together, they have a positive correlation (opens in a new
tab), and this can be seen in the scatter plots, such as in the following plot of human height and
weight data(opens in a new tab). We can see that generally, as height increases, so does weight.
The line shown is the trend line which can be added in Excel by selecting the scatter chart,

If one variable increases as the other decreases, the two variables have a negative correlation,
 D. Histograms

A Histogram is a column chart that measures the frequency of data in a data set and specifically
groups numerical values into bins we define.

Column Charts vs Histogram

Recall that we previously created a column chart to compare counts of categories within a data
set. This kind of chart answers a question like: how many players are there in each playing
position in the league?

But what if we want to ask the question: how many players made under $1 million in salary, and
between $1 and $2 million, and between $2 and $3 million in salary? This kind of chart is called
a histogram, and the groupings we choose such as, 1) all salaries between $1 and $2 million, and
2) salaries between $2 and $3 million, are the bins.

There are two ways to do this in Excel.

Analysis tool pack add-in:

We'll start with a method that works on both Windows and Mac using the histogram tool in the
analysis tool pack add-in. Instructions for loading the analysis tool pack add-in are given in the
Getting Started instructions.

To create the histogram:

Choose data analysis from the data menu on Windows or from the tools menu on Mac. Choose
histogram, which opens a dialog.

For the input range, select the data from the salaries column.

For the bin range, select the bin intervals you've created.

If you have a label at the top of your columns, click labels.

For the output options, select new worksheet and chart output.

Press OK.

Insert chart
Available in Microsoft Excel. The tool pack histogram requires two columns of data. One for
data you want to analyze, and one for bin levels that represent the intervals for the bins. In the
video example, I started at $1 million, then $2 million, et cetera, up to $15 million. When I
created the histogram, the number of values in the salaries lists that are below $1 million will be
in the first bin. The numbers of salaries between $1 and $2 million will be in the second bin, and
so on.

To create the histogram:

Select your data and click insert, recommended charts, and choose the histogram chart.

To configure details about the bins, right-click the horizontal axis of the chart, click format axis
and then click axis options.

The dialog provides options for choosing categorical data like the player positions or automatic
for numerical data. You can specify the number of bins that you might choose to experiment
with a bit. Note: If you choose bins that are too narrow, the result can be noisy. On the other
hand, too few bins will hide details.

As with other charts, the design and layout can be further customized from the design menu
when the chart is selected.

E. Box Plots
A box plot, which in our case is really a box and whisker plot, is the visualization of statistical
spread in a data set of values.

The five numbers summary

A traditional box plot is built using the five numbers summary. The five numbers summary
consists of five values.

Maximum

Minimum

1st quartile

2nd quartile, aka 'median'

3rd quartile
 F. Professional Presentations
Excel and other spreadsheet applications can do a nice job of creating attractive tables and charts.
It's up to us to make them look their absolute best, though. They should:

Be readable

Be interesting

Show the required information

Not show extra information that doesn't matter

When you present data:

Use fonts and layouts appropriately to include and emphasize the elements that matter and
exclude elements that don't.

Ask some questions:

Who is your audience?

Is this a quick overview presentation on a slide with the data backup elsewhere, or is it a written
technical review where more in-depth data should be presented?

Improving charts

For categorical column data, like the counts of the baseball positions, the relative sizes are easier
to see at a glance if the data is sorted by size. This would not be something we would want to do
for sequential data like daily stock values or histogram bins, but it is more readable in this case.
There should be a prominent and descriptive title.