Excel 2007 for Starters: The Missing Manual

9.4. Chart Types

Although there's a lot to be said for simple column chartsthey can illuminate trends in almost any spreadsheetthere's nothing quite as impressive as successfully pulling off the exotic bubble chart. This section covers the wide range of charts that Excel offers. If you can use these specialized chart types when they make sense, you can convey more information, make your point more effectively, and add a little bit of flair to your workbook.

9.4.1. Column

By now, column charts probably seem old hat. But column charts actually come in several different flavors (technically known as subtypes). The main difference between the basic column chart and these subtypes is how they deal with data tables that have multiple series. The quickest way to understand the difference is to look at Figure 9-12, which shows the same data using several different subtypes.

Here's a quick summary of your column chart choices:

• Clustered Column (Figure 9-12, top left). In a clustered column, each value is shown in its own separate column. To form a cluster, the columns are grouped together according to category. The Clustered Column makes it easy to compare data series, but somewhat more difficult to compare categories.

• Stacked Column (Figure 9-12, top right). In a stacked column, Excel displays only one column for each category. To create this column, Excel adds together the values from every series for each category. However, the column is subdivided (and color-coded), so you can see the contribution each series makes. The Stacked Column is an elegant way to compress the data, and it lets you compare series totals without losing individual series information.

• 100% Stacked Column (Figure 9-12, bottom left). The 100% stacked column is similar to a stacked column in that it uses a single bar for each category and subdivides that bar to show the proportion from each series. The difference is that a stacked column always stretches to fill the full height of the chart. That means stacked columns are designed to focus exclusively on the percentage distribution of results, not the total numbers.

• 3-D Effect Column (three types). There's a 3-D version for each of the three basic types of column charts, including clustered, stacked, and 100% stacked. The only difference is that the chart is drawn with a dash of special effects, making the columns look three-dimensional.

• True 3-D Column (Figure 9-12, bottom right). The true 3-D column chart (the last entry in the list of subtypes), shows a bar for each value. However, Excel displays each new series behind the previous series. That means if you have three series, you end up with three layers in your chart. Assuming the chart is tilted just right, you can see all these layers at once, although it's possible that some bars may become obscured, particularly if you have several series.

9.4.2. Bar

The venerable bar chart (Figure 9-13) is the oldest form of data presentation. Invented sometime in the 1700s, it predates the column and pie chart. Bar charts look and behave almost exactly the same as column chartsthe only difference being that their bars stretch horizontally from left to right, unlike column charts, where they rise from bottom to top.

Excel provides almost the same set of subtypes for bar charts as it does for column charts. The only difference is that Excel doesn't offer a true three-dimensional (or layered) bar chart, although it does offer clustered, stacked, and 100% stacked bar charts that give a three-dimensional effect.

9.4.3. Line

Line charts are almost always used to show changes over time. Line charts emphasize trends by connecting each point in a series with a line. The category axis represents a time scale or is just a set of regularly spaced labels.

Excel provides several subtypes for line charts:

• Line. This is the classic line chart, which draws a line connecting all the points in the series. The individual points aren't highlighted.

• Stacked Line. In a stacked line chart, the first series is displayed just as it is in the normal line chart, but the second line consists of the values of the first and second series added together. If you have a third series, it displays the total values of the first three series, and so on. Stacked line charts are sometimes used to track things like a company's cumulative sales (across several different departments or product lines), as Figure 9-14 demonstrates. (Stacked area charts are another alternative, as shown in Figure 9-16.) Stacked line charts aren't nearly as common as stacked bar and column charts.

• 100% Stacked Line. A 100% stacked line chart works the same as a stacked line chart in that it adds the value of each series to the values of all the preceding series. The difference is that the last series always becomes a straight line across the top, and the other lines are scaled accordingly so that they show percentages. The 100% stacked line chart is rarely useful, but if you do use it, you probably want to put totals in the last series.

• Line with Markers (3 types). These subtypes are the same as the three previous line chart subtypes, except they add square boxes to highlight each data point in the series.

• 3-D Effect Line. The 3-D effect draws ordinary lines without markers but adds a little thickness to each line with a 3-D effect.

9.4.4. Pie

Pie charts show the breakdown of a series proportionally, using "slices" of a circle. Pie charts are one of the most recognizable types of charts, and one of the simplest (they work only if you're trying to chart a single data series).

Here are the pie chart subtypes you can choose from:

• Pie. This is the basic pie chart everyone knows and loves, which shows the breakup of a single series of data.

• Exploded Pie. The exploded pie separates each piece of a pie with a small amount of white space. Usually, Excel charting mavens prefer to explode just a single slice of a pie for emphasis. This technique uses the ordinary pie subtype, as explained in Excel: The Missing Manual.

• 3-D Effect Pie (2 types). The pie and exploded pie types can both be rendered in three dimensions, tilted slightly away from the viewer for a more dramatic appearance.

• Pie of Pie. With this subtype, you can break out one slice of a pie into its own, smaller pie (which is itself broken down into slices). This is a great way to emphasize specific data.

• Bar of Pie. The bar-of-pie subtype is almost the same as the pie-of-pie subtype. The only difference is that the breakup of the combined slice is shown in a separate stacked bar, instead of a separate pie.

9.4.5. XY (Scatter)

XY Scatter charts show the relationship between two different sets of numbers. Scatter charts are common in scientific, medical, and statistical spreadsheets. They're particularly useful when you don't want to connect every dot with a straight line. Instead, scatter charts let you use a smooth "best fit" trendline, or omit the line altogether. If you plot multiple series, the chart uses different symbols (like squares, triangles, and circles) for each series, ensuring that you can tell the difference between the points.

Why would you want to plot data points without drawing a line? One reason is that you may need to draw conclusions from an inexact or incomplete set of scientific or statistical data. For example, scientific types may use a scatter chart to determine the relationship between a person's age and his reflex reaction time. The problem is that no matter how disciplined the experimenters, they can't test every different age. In addition, their data will include natural variations from the overall trend. (In other words, if the trend is for older people to have gradually slowing reactions, it's still likely that you'll run across a few exceptionally speedy older folks.) In this case, the best approach is to include no line or use a smooth "best fit" line that indicates the overall trend, as shown in Figure 9-15.

Excel offers several scatter chart subtypes, including:

• Scatter. The scatter chart uses data markers to show where each value falls. It adds no lines.

• Scatter with Smoothed Lines. This scatter chart adds a smooth line that connects all the data points. However, the points are connected in the order they occur in the chart, which isn't necessarily the correct order. A better option is to add a trendline to your chart, as explained in Excel: The Missing Manual.

• Scatter with Straight Lines. This subtype is similar to the scatter chart with smoothed lines, except it draws lines straight from one point to the next. It works like a line chart, and it makes sense only if you have your values in a set order (for example, from lowest to largest or from the earliest date to the latest).

• Scatter with No Data Markers (2 types). These subtypes are identical to the scatter with smoothed lines and the scatter with straight lines. The only difference is that they don't show data markers for each point. Instead, all you see is the line.

9.4.6. Area

An area chart is very similar to a line chart. The difference is that the space between the line and the bottom (category) axis is completely filled in. Because of this difference, the area chart tends to emphasize the sheer magnitude of values rather than their change over time, to the point that it can actually obscure some data series. Figure 9-16 demonstrates.

Area charts exist in all the same flavors as line charts, including stacked and 100% stacked. You can also use subtypes that have a 3-D effect, or you can create a true 3-D chart that layers the series behind one another.

9.4.7. Donut

The donut chart (Figure 9-17) is actually an advanced variation on that other classic food-themed chart, the pie chart. But while a pie chart can accommodate only one series of data, the donut can hold as many series as you want. Excel displays each series in a separate ring. The rings are one inside the other, so they all fit into a single compact circle.

The donut chart is ideal for comparing the breakdown of two different sets of data. But keep in mind that most people emphasize the data on the outside ringit's a human thingso make sure this actually is the most important series. To change the order of the series, you can modify the series in step 2 of the Chart Wizard, as described on Section 9.1.2.

Excel offers two donut subtypes: standard and exploded. (An exploded donut doesn't suggest a guilty snack that's met an untimely demise. Instead, it's a donut chart where the pieces in the topmost ring are slightly separated.)

Although the donut chart can hold as many series as you want, if you add more than two or three, the chart may appear overly complicated. No matter what you do, Excel never fills in the center of the donut.

9.4.8. Radar

The radar chart (Figure 9-18) is a true oddity, and it's typically used only in specialized statistical applications. In a radar chart, each category becomes a spoke, and every spoke radiates out from a center point. Each series has one point on each spoke, and a line connects all the points in the series, forming a closed shape. The effect of all these spokes and lines is to make the chart look something like the radar on an old-time submarine.

There are three radar subtypes: the standard radar chart, a radar chart with data markers indicating each point, and a filled radar, where each series appears as a filled shape, somewhat like an area chart. No matter what subtype you use, choosing data for a radar chart isn't easy.

9.4.9. Surface

A surface chart shows a 3-D surface that looks a little like a topographic map, complete with hills and valleys. Surface charts are different from most other charts in that they show the relationship of three values. There are actually two category axes (x and y), which determine the position of a data point. The value links all the points to create a surface.

Surface charts are neat to look at, but ordinary people almost never create them. They're definitely overkill for tracking your weekly workout sessions. One problem is that to make a good surface chart, you need a lot of data. (The more points you have, the smoother the surface becomes.) Your data points also need to have a clear relationship with both the x-and y-axes (or the surface you create is just a meaningless jumble). Usually, rocket-scientist types use surface charts for highly abstract mathematical and statistical applications. Figure 9-19 shows an example.

9.4.10. Bubble

The bubble chart is an innovative variation on the scatter chart. It plots only a single series, and it never draws a line. Each point is marked with a circleeither an ordinary circle or a three-dimensional sphere (bubble), depending on the subtype you choose. The extra frill is that the bubble sizes change based on a second set of related values. The larger the value, the larger the data-point bubble. Figure 9-20 shows an example.

9.4.11. Stock

A stock chart displays specialized data for stocks. Usually, these charts show how a stock value changes over a series of days. The twist is that the chart can display information about the daytime high and the daytime low of the stock, along with its opening and closing value. Excel uses all this information to draw a vertical bar from the stock's low point to its high point on a given day. If you're really ambitious, you can even add volume information (which records the number of shares traded on a given day).

Stock charts are more rigid than most other chart types. In order to use a stock chart, you need to create a column of numbers for each required value. The type

of columns you need and their order depends on the stock chart subtype that you select. Here are your choices:

• High-Low-Close

• Open-High-Low-Close

• Volume-High-Low-Close

• Volume-Open-High-Low-Close

In each case, the order of terms indicates the order of columns you should use in your chart. For example, if you select Volume-High-Low-Close, the leftmost column should contain the volume information, followed by another column with the stock's daytime high, and so on. (Technically, you can use step 2 of the Chart Wizard to specify each series, even if it's not in the place Excel expects it to be, but it's tricky to get right, so it's easiest to just type in data in the order indicated by the chart type name.) No matter which subtype you use, a stock chart shows only values for a single stock. Figure 9-21 shows an example of a Volume-High-Low-Close chart.

9.4.12. Cylinder, Cone, and Pyramid

The cylinder, cone, and pyramid are really just more exotic versions of the tried-and-true column and bar charts. But instead of using ordinary rectangles, these charts use 3-D cylinder, cone, and pyramid shapes. As with column and bar charts, you can specify how cylinder, cone, and pyramid charts should deal with multiple series. Your options include clustering, stacking, 100% stacking, and layering (true 3-D). Figure 9-22 shows a few examples.