Just how much information should you include in your charts? Charts should contain enough information for the reader to quickly and reasonably understand the ideas that are being displayed. Clear labels and legends should demonstrate what is being measured and in what units. In some cases, readers will expect to know data values at each point within the chart.
Although the ineffective line chart in Figure 12.2 does contain a legend showing which lines correspond to which hotels, the meaning of the y-axis is not as clear. A reader may assume that the data comes from a survey, since ratings is in the title, but be unsure what the range or direction of the scale is. By contrast, the more-effective line chart in Figure 12.2 contains a note indicating the range of the scale. Many charts place this information in a label along the y-axis.
Another basic purpose of a chart is to convey complicated information as quickly as possible. If your readers can’t process the information rapidly, they will lose interest. To some degree, this requires a balancing act with information sufficiency. The more information you provide, the more difficult it may be for some readers to process the chart quickly. By selecting only the necessary information and placing labels and data at appropriate places, you enable your reader to process the information quickly and efficiently. Ideally, your reader should grasp the key ideas within 10 to 15 seconds.
The less-effective line chart in Figure 12.2 reveals several processing problems. The most serious is that the legend forces the reader to glance back and forth between the lines and the legend to correctly link the data series. Another problem is that the Prestigio data series, which should be the center of attention, is placed underneath the other lines, with no special formatting features to make it stand out. The more-effective chart is far easier to process. Data labels appear directly next to each line so that the reader does not have to glance back and forth between the legend and the plot area. Furthermore, the Prestigio line is bolder and thicker, and it is placed in front of the other lines to draw the intended attention.
An effective chart leaves a lasting impression about your key point. Will your readers remember your intended main message in two hours? If not, your chart had little impact. The takeaway is the essence of your chart—how the information, title, focal points, and other formatting combine to convey a lasting message. Overall, the ineffective line chart in Figure 12.2 leaves little lasting impression. The reader who studies the chart carefully might see that the Prestigio’s staff and service ratings improved more than did those of competitors, but the reader has to get through a compilation of colored lines with little or no contextual reference. Furthermore, the chart offers no explanation for why this change in ratings may have occurred. By contrast, a reader can rapidly process the more-effective line chart in Figure 12.2. The title, focal points, and simple design lead to one strong takeaway message: The Prestigio launched a staff and service initiative that has successfully improved customer satisfaction compared with its major competitors. Figures 12.3, 12.4, and 12.5 present other types of charts with less-effective and more-effective variations. Figures 12.6 and 12.7 present a variety of other useful formats for charts.
Figure 12.3 Less-Effective and More-Effective Pie Charts
Key Design and Formatting Problems in Less-Effective Chart Adjustments in More-Effective Chart Title Descriptive but unexciting title. Descriptive title focuses attention on the fact that these are 3-day conference attendees. Focal points The main focal point is the large pie slice. The colors used give a very dense and dark feeling to the visual. The primary focal point is the slice highlighting those not purchasing any Internet service. It is labeled more effectively (“No Purchase of Internet” versus “0 days” in the less-effective chart) and is written in bold text on a darker-colored background to draw attention to this key point. Information sufficiency Absence of data label on each slice makes this chart difficult to interpret. Data labels are provided in percentages. Ease of processing Legend is placed on the bottom. This forces the reader to move back and forth between the legend and the pie slices in the plot area. Also, the breakaway, 3-D shape of the object skews the data. The pie slices are not arranged for fastest processing. Data series names and data labels are placed together in the pie slices to foster easy processing. The largest pie slice is located at 12 o’clock for quick recognition (most people read pie charts beginning at 12 and continue to read in a clockwise direction). Takeaway message Most conference attendees do not purchase Internet services. However, getting the message requires a great deal of effort and could easily be missed or forgotten quickly. All aspects of the chart collectively demonstrate that conference attendees are unlikely to purchase Internet services.
Figure 12.4 Less-Effective and More-Effective Bar Charts
Key Design and Formatting Problems in Less-Effective Chart Adjustments in More-Effective Chart Title descriptiveness Nondescriptive, bland title. Title immediately recognizes the Prestigio’s leading position in dining ratings. Focal points Lacks focal points. All bars are treated equally. Darker color of the Prestigio bar draws attention to it. Information sufficiency Inadequate information about the rating scale. A note about the rating scale and inclusion of data labels provides sufficient information. Ease of processing The legend is unnecessary and distracting. The items are not ordered effectively (the order is neither alphabetical nor quantitative) to help draw rapid comparisons. The large gap size compared to bar width reduces quick processing. The axis increments are in rarely used units (generally, units in multiples of 2, 5, and 10 are more natural). The chart is arranged in descending order by average ratings to make comparisons easier. Bar width in comparison to gap width is most conducive to rapid processing. Takeaway message The takeaway message is that the Prestigio has higher dining ratings. However, the message is weak and could easily be glossed over or forgotten. The Prestigio occupies the proud position of leading its competitors in dining ratings. This is a strong, optimistic, and memorable message.
Figure 12.5 Ineffective Clustered-Column Chart and More-Effective Panel of Charts
Ineffective Clustered-Column Chart Effective Alternative: Panel of Charts Title Descriptive but bland. Curiosity building (“How the Prestigio Stacks Up”); a call to action (“Room for Improvement in…”). Focal points None. Too cluttered. Prestigio rankings and position for each rating area. Information sufficiency No data labels. Data labels provided for each rating area. Ease of processing Nearly impossible. Too much information. Not sorted. Simple and easy processing for each rating area. Charts are organized by relative performance (excellent performance on left side, needs improvement performance on right side). Takeaway message No key point related to the ratings. The Prestigio is elite in various areas compared to its competitors, but is behind in other key areas.
How the Prestigio Stacks Up
Room for Improvement in Cleanliness, Meeting Rooms, Business Center, Staff & Service
Note: Ratings are on a scale from 1, poor, to 5, excellent. All ratings were retrieved from the Wahoo travel website and are averaged across the year.
Doughnut charts allow you to represent wholes. Unlike pie charts, you can present more than one data series.
High-low charts allow you to show values that fluctuate. These charts are often used for stock prices.
Histograms allow you to represent frequencies. Frequencies often reveal data relationships not easily visible by looking at averages.
Scatter plot or X-Y charts allow you to include pairs of data on an x-y plot. Many scatter plots contain trend lines to reveal data relationships.
Org charts allow you to show who various personnel report to.
Gantt charts allow you to show progress on aspects of projects. These charts are frequently used as part of project management.
Flowcharts allow you to depict a series of steps in a process or procedure to help others make decisions.
Chevron lists or charts allow you to show a set of sequential steps and provide subpoints for each step.
Although formatting a chart is secondary to creating a powerful takeaway message, it is by no means unimportant. Since visuals have an impact even before the reader begins reading, ineffective formatting can give the reader an impression of sloppy or imprecise work.
Generally, the formatting should be as simple as possible and should accentuate the key data relationships. If a formatting feature detracts from the key points, remove or improve it. Table 12.6 provides general formatting guidelines for charts.
|Chart Type||Formatting Guidelines|
Generally, charts are the most effective way of quickly demonstrating a key point or relationship. However, charts are limited in the amount of information they can provide.
Tables, by contrast, allow you to provide more data with additional precision. Because of this, charts are generally better for highlighting a key idea, and tables are generally better for comprehensiveness and precision.
Like charts, tables are typically more effective with simple formatting. In addition, the way a table presents data can affect the clarity of its message. Consider, for example, the tables in Figure 12.8, which are based on identical data. Place yourself in the position of the reader and assume you have the following question: “Does higher income level correspond with higher likelihood of purchasing Internet services?” It is difficult to answer this question quickly by looking at the less-effective table. By contrast, glancing at the more-effective table rapidly reveals that purchasing no Internet service (0 days) strongly correlates with the lowest income bracket (under $30,000/year).
Figure 12.8 A Less-Effective and More-Effective Table
During the three days of the conference you attended at the Prestigio, how many days did you purchase Internet service? Days of Internet Service 0 1 2 3 All Respondents 154 15 31 36 Gender Male 82 8 15 22 Female 72 7 16 14 Income Under $30,000 15 0 1 2 $30,000–$40,000 41 4 3 7 $40,000–$50,000 48 3 11 12 $50,000–$75,000 33 6 7 8 $75,000–$100,000 12 2 4 4 Over $100,000 5 0 5 3
Internet Service Purchases among Conference Guests
Days of Internet Service Purchased (Number of Respondents in Parentheses) 0 Days 1 Day 2 Days 3 Days Total (#) All Respondents 65.5% (154) 6.4% (15) 13.2% (31) 15.3% (36) 236 Gender Male 64.6% (82) 6.3% (8) 11.8% (15) 17.3% (22) 127 Female 66.1% (72) 6.4% (7) 14.7% (16) 12.8% (14) 109 Income Under $30,000 83.3% (15) 0.0% (0) 5.6% (1) 11.1% (2) 18 $30,000–$40,000 74.5% (41) 7.3% (4) 5.5% (3) 12.7% (7) 55 $40,000–$50,000 64.9% (48) 4.1% (3) 14.9% (11) 16.2% (12) 74 $50,000–$75,000 61.1% (33) 11.1% (6) 13.0% (7) 14.8% (8) 54 $75,000–$100,000 54.5% (12) 9.1% (2) 18.2% (4) 18.2% (4) 22 Over $100,000 38.5% (5) 0.0% (0) 38.5% (5) 23.1% (3) 13
The less-effective table is cluttered due to excessive grid lines, poor labels, and non-indented items. By contrast, the more-effective table limits the number of grid lines. Furthermore, each grid line serves a distinct purpose. The initial grid lines separate the column labels from the survey data. Subsequent grid lines separate each category of data, including those for all respondents, gender, and income level. Indents of items within each category further accentuate the distinctions between categories.
The second table also is more effective because numerical adjustments have been made. The first table contains counts of respondents who responded in certain ways. Counts make it difficult for readers to make effective comparisons quickly. Yet, many readers are also interested in knowing how many people participated in a survey. By converting the counts into percentages, the more-effective table enables readers to process the information more easily. Placing the counts in parentheses makes the data comprehensive.