One of the more important methods of communicating measurement results is by using charts and graphs to visually depict data. The adage that a picture is worth a thousand words holds just as true for data visualization as it does for a photo—the impact of a chart with a sales line swooping up is greater than reading through a paragraph full of purchasing data.
Images and descriptions each have their place, and the ability to use charts to effectively tell a story along with a narrative is all part of being an effective communicator. So how should you approach using charts and graphs? When are the default charts in a monitoring tool sufficient, and when should you export data to create your own charts?
Keep the receiving audience in mind
As you pull together a report and are considering using charts and graphs, your most important task is to keep your reading audience top of mind. If they are familiar with the content and have expectations on how information will be presented, it’s far easier to decide what to include, how to frame the content, and the types of charts that will be used.
However, if your reading audience is new to the data and topics you are including, you need to use care when designing your charts. A typical problem is the chart with too much data included in it, which has the potential to overwhelm and confuse the reader. When deciding how to visually depict your data, consider the following:
- Consider outlining your report first. This will allow you to map out the narrative you are hoping to convey. Constructing an outline makes it easier to stay on track with the report. For example, if you are pulling together an annual media and messaging analysis, you will be following a specific chronology, and layering in how messaging either did or did not resonate during the course of the year.
- What is the primary message or data point that you are trying to convey? If it is a comparison chart, such as showing prior coverage to current year’s, make the comparison a simple one. If there isn’t a need to include other data, such as sentiment rating, in this chart, resist the urge to do so—it clutters the message. You can include sentiment data in another chart, following the narrative you’ve mapped out. Each chart or graph included should be able to visually depict your point at a glance. If it doesn’t, there is probably either too much information in the chart, or you might have selected the wrong style of chart to depict your data set.
- Keep colors, graphics, imaging, and symbols simple. A quick test to determine how effective a graph is for a reader is how many times the reader has to refer back to a legend or key to figure out what is going on in the graph. If there’s too much going on, the reader has to try and process the legend first, and then the graph itself. If there is a lot going on, consider how you could break the information down into more digestible bits. For example, rather than including all media types in one graph, see if there’s an effective way to break out by source type: major newspapers, television, and blogs. This is also a way to tackle data that represent small percentages. If you are trying to graph data that have four or five primary items and then a bunch of smaller items that make up the remaining, instead of having a chart that includes all, consider using a pie chart with a “callout” chart (sometimes called a “pie of pie” chart). This allows you to include all of the relevant information while preserving readability.
- Make certain to include all applicable data. In your efforts to maintain simplicity, don’t go overboard and eliminate information that is necessary to understand the chart. Make sure that axes, pie segments, and bars are labeled, that the chart is correctly scaled.
- Use a variety of graphing types, considering carefully how best to depict each data set. Using nothing but pie charts throughout a report will feel repetitive, so pay attention to the types of charts you choose. To show change over time, it is generally best to consider an x/y axis chart. Pie charts are best for showing percentages of a whole, such as what media types were included in the analysis. “Doughnut” charts also are good for percentages of a whole, so they can often be used interchangeably with pie charts.
Considerations for presentations
Using charts and graphs in reports that are designed to be read through allow for more detail to be conveyed by the nature of the format. Interspersing charts with text provides a flow for the reader: background and context, chart image, followed by an explanation of how the chart fits into the overall narrative of the report.
If you are developing charts and graphs for a presentation, the narrative will need to be distilled into very short bullet points and the charts need to be a simple and elegant as possible. This is particularly true if you’ll be presenting to a big room, as crowded graphs and multiple bullet points up on screen will have people squinting to figure out the presentation slide, rather than listening to you speak.
Keep visuals informative, but reduce clutter and distractions for presentations. Make certain that fonts are easily readable, and use logical abbreviations when possible; instead of spelling out “third quarter,” use Q3—it conveys the same information in a much shorter format.
Using charts and graphs can significantly enhance the ability of your audience to digest measurement data. Spending just a bit of time planning out what you wish to convey in a report or in a presentation allows you to construct a narrative that will flow logically for your audience, and most effectively convey the measurement data you have collected and analyzed.