Properties of the Scale

Notice that in Figure 9-1 the mapping of the data by the scale is defined by a specific property listed under the scale, for example, “Shape1.Fill” or “Shape1.Height”. Each scale will have listed under it similar properties that identify each attribute to which a field has been bound and whose data is mapped by that scale. In Figure 9-2, we have set out three examples of how these properties may appear in the Scales pane.

In Example #1, two different categorical fields have been bound to attributes generating two color scales. In “Scale1,” a categorical field has been bound to both the Fill and Stroke attributes of a shape. In “Scale2,” a different categorical field has been bound to the Fill and Stroke attributes of a symbol. Each scale maps a different collection of colors to the items in each categorical field.

In Example #2, the same numerical field or three different numerical fields have been bound to three different attributes. Either scenario would result in the generation of three separate numerical scales. Binding a numerical field to a different numerical attribute always generates a new scale, even if the field bound to the attribute is the same.

In Example #3, the same numerical field or different numerical fields have been bound to the same attribute of two different shapes and two different symbols. With respect to the rectangles, this could be because either the chart contains a glyph comprising two rectangle shapes or the chart contains two separate glyphs comprising a single rectangle shape (we will be using multiple glyphs in Chapter 11). This same arrangement could also apply to the two symbols. In both cases, each field that is bound to the same attribute will share the same scale. We take a detailed look at this scenario in the section on “Numerical Scales” below.

What you can infer from the preceding three examples is that color scales are specific to the categorical field that is bound to attributes (and can also be specific to numerical fields that are bound to nonnumerical attributes; see the section on “Color Scales” below). Numerical scales, on the other hand, are specific to the attribute whether the same or a different numerical field has been bound.

Perhaps, you will also appreciate at this stage that numerical scales will be more challenging to understand, so let’s start with something a little simpler and look more closely at Charticulator’s color scales.

Color Scales

A color scale is created when you bind a field to an attribute that has a color associated with it such as Fill, Stroke (for borders), or Color (for text marks). If you bind a categorical field to one of these attributes, Charticulator will map different colors to each member of the category, creating the associated color scale. If you bind a numerical field to the attribute, this will create a gradient color scale to define high and low values; see Figure 9-3.

We looked briefly at editing color scales in Chapter 3, so now let’s just recap on what we have learned regarding how color scales can be managed.

Editing Color Scales

You can edit the colors of the scale either by clicking the attribute or by clicking the property name in the Scales pane that defines the color you want to edit. For instance, if you click on the “Shape1.Fill” property in the Scales pane, a dialog will open to allow you to edit the colors used for either the fill of the rectangle shape for a categorical field such as “Salespeople” or for the gradient color used by a numerical field such as “Sales”; see Figure 9-4.

To edit the categorical colors, click a color and then reselect a color from the color palette. To edit a gradient scale, select from the Palettes tab or use the Custom tab to create your own custom gradients.

It’s now time to turn our attention to the more problematic numerical scale.

Numerical Scales

A numerical scale is created when you bind a numerical field to a numerical attribute such as Height, Width, Size, or X Span/Y Span. For example, in Figure 9-5, we have bound the “Sales” field to the Height attribute of a rectangle shape and to the Size attribute of a symbol.

One of the most important things to understand when working with Charticulator’s numerical scales is that once Charticulator has generated a numerical scale to map a numerical field to a numerical attribute, that scale will be used for all subsequent mappings to that same attribute. Each mapping is listed as a property under the scale that has been set by the first bound field. For example, under “Scale1,” there may be two properties, “Shape1.Height” and “Shape2.Height,” where two different numerical fields have been bound to the Height attribute of two different rectangles that comprise the glyph (we will be doing this presently). When the second numerical field was bound to the Height attribute of the second rectangle, the scale had already been set by the first field that was bound.Therefore, the order that you bind the numerical fields is critical. Not understanding the impact of the order in which numerical fields are bound can result in unexpected behaviors of the glyph.

Consider the chart in Figure 9-6 where we are plotting two numerical fields, “2019 sales” and “2020 sales,” that have been bound to the Height attribute of two rectangle marks.

You can see that the rectangle mark for “2019 sales” is plotted correctly, but the rectangle mark for “2020 sales” spills outside the plot segment.

Why is the rectangle mark for “2020 sales” exhibiting this bizarre behavior? Before we answer this question, let’s get to grips with how the glyph comprising the two rectangles was generated.

The building of this glyph was a little more challenging than simply adding two rectangles to the Glyph pane and then dragging the numerical fields into the Height attribute, so we have set out in Figure 9-7 how the glyph was constructed.

To start, the first rectangle was drawn inside the guides of the Glyph pane, ensuring that the rectangle was anchored to the top, vertical middle, side, and bottom guides of the glyph. The intuitive action to take next was to bind the first numerical field, that is, “2019 sales,” to the Height attribute of this shape generating “Scale1” in the Scales pane.

The second rectangle was drawn starting above the top guide of the glyph and then anchored to the vertical middle, side, and bottom guides. “2020 sales” was then bound to the Height attribute of the second rectangle, but this resulted in the rectangle spilling outside the plot segment.

The reason that the rectangle for “2020 sales” is too tall for the plot segment is that we bound the “2019 sales” field first to the Height attribute and therefore set the Domain Start and End values for “Scale1”, mapping the height of the rectangles to “2019 sales.” Because the maximum value for “2019 sales” is 368127, this set the Domain End value of “Scale1” to 370000. See Figure 9-8.

If you were to insert a numerical legend for “2019 sales,” this would also reflect this scale. When we then bound “2020 sales” to the Height attribute of the second rectangle, this rectangle was also scaled according to “Scale1”. This in itself isn’t necessarily the problem. It’s the fact that the maximum value for the “2020 sales” is 755892 which is greater than the Domain End value of “Scale1”, and what you see in Figure 9-6 is the result of this mismatch. What we should have done was to bind “2020 sales” to the Height attribute first, setting the correct scale for both rectangles. This would also entail drawing the rectangles in the glyph in a different order as shown in Figure 9-9.

Doing this will generate the correct Domain End value that can be used for both numerical fields (see Figure 9-10), and both rectangles will be plotted correctly.

The takeaway from this section is not the specifics of building the clustered column chart using the two numerical fields, which you may feel at this stage is overly complicated. The takeaway is understanding how Charticulator calculates numerical scales and its implications on how the data is plotted on the chart: that binding the first numerical field always sets the numerical scale. You will have another shot at building a similar clustered column chart in Chapter 14 when we explore working with multiple measures, and we’ll find easier ways to construct it correctly.

Editing Numerical Scales

Just as with color scales, you can edit a numerical scale either by clicking the attribute or by clicking the property in the Scales pane, for example, by clicking “Shape1.Height”. When you edit a numerical scale, you have a choice of editing the Domain and the Range attributes; see Figure 9-11.

We look at editing the Domain attribute in the section on legends (see the section on “Editing the Scale of a Numerical Legend” below), but for now, let’s focus on what the Range attribute can do for us.

Well, in fact, if you’re working with a mark or line as opposed to a symbol, changing the Range attribute won’t do that much for you. The Range attribute of the scale determines the difference between the smallest glyph and the biggest. In our chart, this is the difference in height between the shortest rectangle and the tallest. If you’re using a mark or a line, when you first bind numerical data, Charticulator will automatically scale the height, width, or span of the mark or line to fit the plot segment. This means that if you change the size of the plot segment, the glyph automatically rescales. Therefore, if you put measurements into the Start and End values of the Range, you will constrain the size of the glyph; see Figure 9-12.

You may find more reason to change the values in the Range attribute if you’re using a symbol as your glyph and have a numerical field bound to the Size attribute. Editing the Range values allows you to resize the symbol proportionally. When you click the “Symbol1.Size” property in the Scales pane, you’ll see that unlike the rectangle mark, the symbol has a specific End value in the Range attribute (628.32), constraining the size of the symbols irrespective of the size of the plot segment. If you edit the End value however, you will increase the relative sizes of the symbols; see Figure 9-13.

Another example of using the End value of the Range attribute to control relative sizes is when you bind a numerical field to the Size attribute of a text mark. This will ensure that the text remains legible for small values but is never too big for the larger ones.

Creating Additional Scales When Mapping Data

You have already learned that when you bind the first numerical field to the Height attribute of a rectangle, this will set the height scale for all subsequent mappings to that attribute. The properties generated under the scale that map the values to the height of the rectangles are specific to the Height attribute rather than to the field, so different fields will always share the same numerical scale for heights. But what if your numerical fields hold values that require different scales to be used by the Height attribute? For example, you may want to compare sales and quantity in the heights of two rectangles in a clustered column chart.

With regard to categorical fields, there will only be one color scale generated for each categorical field, regardless as to which attribute it is bound. But what if you want different colors being used for each of your categorical items when bound to different attributes?

You can see in Figure 9-14 that we have resolved these problems by generating new scales. In the top chart, we have used a second scale for the Fill attribute of the symbol so that we can use a color scale to “group” the salespeople, either red or green, according to their team (the colors mapped don’t have to be different for each categorical item). In the chart underneath, we have used a second scale for the Height attribute of the second rectangle that is plotting the “NoOfSales” numerical field.

To generate a new scale in this way when binding a field to an attribute, you must hold down the SHIFT key as you drag the field into the attribute, as shown in Figure 9-15.

When you do this, you will see a new scale has been created in the Scales pane. New numerical scales will be generated accordingly, but with color scales, you can now edit the colors used by the new scale and use the same color for different categorical items if required.

Reusing Scales

As well as being able to create new scales, you can reuse existing scales by dragging and dropping them into an attribute. For instance, to use the same colors for the borders of a rectangle as you have for the fill color, you can drag and drop the “Shape1.Fill” property into the Stroke attribute; see Figure 9-16.

We’ve been focusing on Charticulator scales, learning how they are generated and how to manage them. However, once a scale has been created, you’ll want a way to visualize the values represented by the scale. Specifically, you’ll need an explanation of the height or width of a mark or the colors that represent the categories or gradients. This is where you’ll need to add Charticulator’s legends to your chart, so let’s move forward and take a closer look at how legends are created and edited.

Creating Legends

There are several ways to create legends. The most intuitive way is to use the Legend button on the toolbar, and this is the only way that you can create a column names legend. This button lists all your fields, and you can select the field for which you want a legend. You can hold down the CTRL key to select multiple column names for the column names legend.

However, there are several drawbacks to using this method to add column values legends. The list of fields on the button dropdown doesn’t depend on the binding of data to an attribute, so it provides only “default” legends. Because of this, it’s ambiguous as to which scale the legend will be mapped. For instance, you may need two legends for the “Sales” field, one for a gradient color scale and one for a numerical scale, but you can only insert a legend for the numerical scale from the toolbar button.

A better way to create column values legends is to use the Scales pane. The advantage of using the Scales pane, as we have seen earlier, is that it’s here that you can also edit the colors that are mapped to the scale or the behavior of a numerical scale. If you click a property under a scale in the Scales pane, for example, “Shape1.Height,” this will open up a Scales dialog from where you can add the Legend.

If you prefer, you can also click the attribute where the data has been bound and that has generated the scale, for example, on the Height attribute.

To place a numerical legend on the x-axis for the Width attribute, you must either click the property in the Scales pane (see Figure 9-18) or click the Width attribute. You can’t use the toolbar button.

Using the Scales pane to generate legends in this way therefore gives you the option to both edit the scale and add the legend. Let’s now explore the differences in the two legend types: column values and column names.

Column Values Legends

When you create a column values legend in Charticulator, it’s listed in the Scales pane along with the scale whose data it maps.

All these legends are visual representations of the data that has been mapped by the scales.

Column Names Legends

The column names legend is used when you have plotted multiple numerical values in your chart. We had an example of this in the chart in Figure 9-10 where we plotted data for “2019 sales” and “2020 sales,” both of which had been bound to the Height attribute of two rectangles comprising the glyph.

To add a legend that describes the colors used by each numerical field represented by the rectangles, you need to select the Column names option on the Legend dropdown and select the fields to be displayed in the legend. Hold down the CTRL key to select the required fields. The legend that is placed on the chart will use the default colors from the Power BI theme, and so you will need to manually map the fill colors of the rectangles that represent each numerical field yourself. Use the Attributes pane of the legend and click the Edit scale colors button. See Figure 9-19.

Note how by creating a column names legend this will generate a new scale in the Scales pane that maps the colors to each column name.

Using the column names legend in this way will be important to you when we explore Charticulator’s data axis in Chapter 14 and we understand the difference between plotting “wide” and “narrow” data.

Formatting and Moving Legends

All your legends are listed in the Layers pane under the Chart group. You can then click a legend name to open the Attributes pane. From here, you can reformat the legend, change the order of the items (for categorical legends), and change its position; see Figure 9-20.

In the examples in Figure 9-20, we have moved the categorical legend to the top of the plot segment and changed the shape to a rectangle.

In the numerical legend, the tick color has been colored green and the tick line color changed to red. The font size has also been increased.

Editing the Scale of a Numerical Legend

Click on the property under the scale, e.g. “Shape1.Height” to edit the scale used by a legend. The Domain attribute of the scale shows the range of values in the numerical field to which the numerical legend is mapped. For example, our data is mapped from 0 to 800,000 which is the maximum value in the “Sales” field (rounded up). If you want to change the Start value of the legend, edit the Start attribute under the Domain attribute of the scale; see Figure 9-21.

Changing the End value of the legend is a little more problematic. By default, your chart will take up the entire space of the plot segment, the largest or widest shape extending to its edge. This is controlled by the automatic scaling in the Range attribute of the scale. If you want more space between the edge of the plot segment and the tallest or widest shape, you can enter an End value in the Range attribute which indicates how tall or wide you want the tallest or widest shape. This will force the legend to increase its values accordingly; see Figure 9-22.

But as mentioned earlier when we were learning about scales, entering a value into the End option of the Range attribute will restrain the tallest shape when you resize the plot segment and generally renders the chart unpredictable.

If you select the Legend in the Layers pane, in the Attributes pane you can format the tick marks and change the number format as shown in Chapter 8.

Learning how to edit the scale of a numerical legend ends our journey through Charticulator’s scales and legends. I hope that now you’ve read through this chapter, the Scales pane inside Charticulator will no longer be a mystery to you. I certainly found it one of the more unfathomable parts of the Charticulator screen when I first started out. However, you will now understand the behavior of scales, that once a scale is generated for an attribute, it doesn’t change unless you hold down the SHIFT key when binding the field. You now also know what the Domain and Range attributes of the Scale are used for and how they are yet another important factor in controlling how the glyphs behave in the chart. In fact, you now know everything that’s important to know about Charticulator’s scales and legends, and this knowledge will be important to you as you move forward.

However, there is just one further mandatory skill you need to acquire before you can move on to tackle the construction of more challenging charts. I’m alluding to the practicalities of laying out elements on the chart canvas and in the Glyph pane. For this, you need to be competent in using Charticulator’s guides and guide coordinators and anchoring elements to them. If you’ve ever tried to move a categorical legend or a chart title, you may have noticed that it has a habit of wandering around the chart without your permission. If you want to remedy this sad situation, read on to the next chapter.