5 Tips for visualizing data with charts
In my recent TAA webinar, “Show Me! The Art of Using Visual Elements to Enhance a Manuscript“, I shared best practices for incorporating tables, figures, and charts into your manuscripts and the tools available for developing those visual elements.
One of my favorite forms of visual element is a chart. Charts combine the visual appeal of figures with the data content of tables and can be quite effective in conveying context and purpose when used properly. For greatest success, I offer the following five tips for chart usage in your next manuscript.
Accurately represent the data set
Regardless of the chart type you use, select options for series, labels, and scale that make it easy for the reader to clearly identify the values shown. Although 3-D effects add a visual appeal to the chart object, they often reduce the accuracy of the image, especially in the case of pie charts (shown below) where the third dimension can make slices in the foreground appear larger than they actually are.
Use an appropriate chart type
Chart types matter. When charting a table of related data values, it is common to use column, bar, or line charts. While column and bar charts are easily interchangeable with the same data set, be selective when using a line chart as a substitute. Line charts connect the data points giving the visual perception that continuity exists between the points. For distinct values that are not related to one another, this impression should be avoided by using columns or bars as disconnected data point representations.
When charting a single data series of values, particularly one where the percentage of contribution of a single data point to the whole is of interest to the reader, a pie chart is more effective. Pie charts visually display slices of proportionate size for the individual data points within the set. As a result, they are not effective when two or more series are being visualized.
Clearly identify key values
Draw attention to the values of significance in your charts. In some cases, all values are equally significant, and an accurate scale or data labels are effective in clearly identifying each data point value. In other cases, a single data point or two may warrant special formatting using coloring or being separated from the rest of the chart visually (as can be done by exploding a piece of a pie chart).
Only chart what is necessary to convey the intended message. Data labels can be effective for clearly identifying key values, as discussed above, but can quickly become clutter in a column chart with a lot of data points. Similarly, if the scales of multiple series are significantly different, as shown below, one series may not even be displayed in the chart and the inclusion of the series in the legend can simply add confusion to the reader. Either move the second series to a new chart, to a secondary axis, or remove it altogether to reduce the clutter.
Ensure that the chart can stand on its own
Whenever you add a chart to your manuscript, consider that the reader may view it independent of the supporting narrative. This is often the case when journals publish charts and figures in appendices rather than inline within the article. For those readers skimming the illustrative elements or in cases where the chart is separated from the narrative, you want to ensure that the meaning and purpose are not lost.
Adding data labels to the chart can associate the actual value of a data point with the visual representation of the data. Data tables included with the chart can also connect the underlying data to the visualization without relying on another source document for the details. And finally, always include a descriptive title that identifies the purpose of the visualization.
Remember that charts can be very effective communication tools for illustrating data in your articles and books. Applying these five tips can ensure that the message you deliver is that which you intended.