Viewing Grouped Data with Stacked Charts
Stacked charts—whether column, bar, or area charts—have features similar to their unstacked counterparts. An obvious difference is that the data displayed in a stacked chart is shown cumulatively.
With stacked charts, you can set the transparency of the data series. The transparency slider shown in Figure 5.9 applies to all the data series. The series cannot be individually set.
Figure 5.9 Adjusting the transparency for your data series.
Transparency plays a more important role with a regular area chart than it does with a stacked chart because valuable data can be easily obscured with a regular area chart (see the lower-left corner of Figure 5.10).
Figure 5.10 Various ways to represent data by using the Area Chart and Stacked Area Chart components.
Xcelsius 2008 has the Stacked Area Chart component, but there is no option to automatically represent data based on its relative contribution, as shown in the top-right corner of Figure 5.10. To do this, you need to prepare your spreadsheet data so that the data is represented in terms of its relative contribution. Mathematically, this is straightforward. In the current example, you simply divide each of the values for the department by the total quantity for the quarter. Because the quantities for the quarters add up to 100%, the maximum limit for all the quarters is the fixed value 1. This is what gives this kind of stacked area chart a horizontal plateau.
You still need to make a further adjustment to the scaling for your Y-axis. Use of auto-scaling will push the maximum value on the Y-axis to a number greater than 1, to something like 1.2. To regain control, you need to set your scale to manual and peg the minimum and maximum values to 0 and 1, respectively (see Figure 5.11). In addition, you have the choice of setting the number of divisions along the Y-axis or the size of the division. Both of these approaches are equally suited because your scale is fixed.
Figure 5.11 Setting the chart scaling for displaying relative contribution.
Avoiding Needless Data Series Congestion
One of the challenges of presenting information in a dashboard setting is that graphical displays can easily get overcrowded with data. The List Builder component allows a dashboard user to cope with this situation by enabling him or her to select which data series to display and in which order.
The Stacked Area Chart component is useful, but it is even more useful when combined with other components, such as the List Builder component (see Figure 5.12).
Figure 5.12 A List Builder component lets you choose which data series to plot on a display.
In the bottom-right corner of Figure 5.12, the data is displayed in tabular form, using the List View component. List View components have several benefits:
- They are scrollable.
- The column widths are individually adjustable.
- You can sort the data on any column by clicking the appropriate column header. You can toggle between sorting in ascending order and sorting in descending order.
List Builder components are great for when you want to cherry-pick certain pieces of information. Say, for instance, that you want to compare the sales performance of two managers.
There are circumstances in which you may want to see the totality of all the data but want to lump the smaller data values into one big group. You might, for example, be analyzing sales and want to see the detail for your four or five biggest customers and also see the combined total of all the remaining customers. Having a slider to magically set the dividing line between showing details and grouping the remainder would be very convenient.
Dynamically grouping or lumping data is especially important when it comes to Pie Chart components. Figure 5.13 shows an example of this.
Figure 5.13 Dynamic data grouping lets you control how much detail you want to see.
The framework for implementing dynamic data grouping is straightforward. In your underlying spreadsheet, start by having your data sorted from largest to smallest (see column C in Figure 5.14).
Figure 5.14 Populate columns F and G with only the data needed based on the input cell.
Place an input cell (see cell C4 in Figure 5.14) whose value is set by a slider or some other selector-style component, such as a Dial or Spinner control.
Calculate the total amount of sales or whatever you are displaying for the top-tier customers or items displayed in your Pie Chart component (see cell C2 in Figure 5.14). This is based on the value in the input cell, as set by your slider- or selector-style component. In this example, subtract the top-tier sales from the total sales to get the sales for “all others.”
Populate a portion of the spreadsheet (such as columns F and G) with information needed for the Pie Chart component. You can use a formula like this:
where cell C4 is the location of the input cell that is set by the slider.
Drilling Down with Pie Charts
What good is having lots of data if you can’t get to the underlying details? The quantity of sales in the previous example may be annual sales, which is composed of monthly data. It would be convenient to examine the breakdown of sales on a month-by-month basis. This is accomplished using the drill down feature that is built into many of the Xcelsius 2008 visual components.
In your Pie Chart component’s properties panel, go to the Drill Down subtab of the Behaviors tab. Click the Enable Drill Down check box (see Figure 5.15). You need to specify whether you want to drill down based on position, value, row, column, or status list. You need to specify a destination range and, depending on the type of drill down, a source range. For this example, you want to choose the position—that is, which slice of the pie you want to examine—so it is not necessary to specify the source range.
Figure 5.15 Setting drill down options.
You also need to tell Xcelsius whether you want to drill down whenever the mouse passes over a slice in the Pie Chart component or when the slice is clicked.
Because you only need to find out which slice is selected for drill down, the destination range is a single cell, namely the position. It would be a good idea to set the location for this nearby the input cell set by the slider (in this example, cell C5).
To get the drill down data, it’s just a matter of extracting the particular row from the monthly data based on the input cell (C5 in this example). You could display the retrieved data on a Bar Chart component.
There’s just one problem: If your tabular data follows a left-to-right chronologic sequence (such as January, February, March, and so on), the bar chart displays the most recent month at the top. The result is a sequence of dates that reads downward as December, November, October, and so on. To have the bar chart show a January, February, March, and so on sequence, you need to reverse the retrieved data (see Figure 5.16).
Figure 5.16 Notice that the extracted data needs to go from right to left.
When this is corrected, the dashboard renders as expected (see Figure 5.17).
Figure 5.17 A Pie Chart component with drill down to the monthly data.
As a little extra added touch, you can create a miniature isolated slice, as shown in the upper-right inset of the bar chart in Figure 5.18. This helps provide feedback on what slice of the pie chart is being revealed in detail within the bar chart.
Figure 5.18 XY charts displaying representative demographic data.