Powerful Visualization Choices (part 1)
A common question we are often asked is how to choose the right kind of chart or graph to answer your questions. We know that transforming your data into an effective visualization (any kind of chart or graph) helps you make use of your data. Below are some best practice recommendations and examples to help you select the best kind of chart for the type of data you are analyzing and the questions you want to answer. The most powerful story combines different visualizations in meaningful ways and answers the “so what?”. Related charts, maps and filters can all work together to tell the full story and provide actionable insight to support informed decision-making.
- Bar Chart
Bar charts are one of the most common ways to visualize data. Using a bar chart allows you to quickly compare information and see highs and lows at a glance. They are best used when you want to show comparisons between categories. These can be plotted either horizontally or vertically.
When to use:
- To compare data across categories
Make it shine:
- Add color to bars for more impact. For example, if your bar chart show the number of events per year, you can color it to show the number of registrations. In this way you can easily see if the number of events per year is roughly (but not always) correlated to increasing registrations.
- Use stacked bars or side-by-side bars. To add additional depth to your analysis, try displaying related data on top of or next to each other. This can allow you to address multiple questions all in one chart.
- Use as a filter. You can set the bar chart as a filter so when you click on a bar, other charts or maps (which may be viewed together on a dashboard) update to include only the selected category.
- Use vertical column charts to compare a few items; use horizontal bar charts to compare many items.
- Number of events each year, registrations with different registration types, website traffic by primary topic, percent of revenue by product type
Figure 1: Bar code shading. In this bar chart, you can see that we are showing the number of events per year with the height of the bars. We can quickly see that 2014 has the most events. By adding color for the number of registrants, we see something interesting emerge. Look at 2014. It has the highest number of registrants, even though it does not have the highest number of events.
- Line Cart
The line chart is another frequently used chart type in which most people are familiar. A line chart connects individual numeric data points that allows the viewer to easily see the sequence of values. Using a line chart allows you to display trends over a period of time.
When to use:
- Show trends over time
Make it shine:
- Combine a line graph with a bar chart.
- Use shading. When you have two or more line charts, you can fill the space under the respective lines to create an area chart. This allows the view to see the relative contribution that the line contributes to the whole. For example, you might have a line chart showing the primary reason a member joined for the past several years. By making it into an area chart, you can easily see how many new members you’ve had over time and the effectiveness of certain campaigns.
- Include a trend line. Using a trend line or goal line can show you how you are doing against a benchmark.
- Membership counts over a five-year period, website page views during a quarter, meeting revenue for the past 3 years
Figure 2: Simple Line Graph with Trend Lines. This is a simple line graph showing the number of registrations each month leading up to the annual conference in October. The line in green shows 2015. It looks like we are in for a good year!
- Pie Chart
The pie chart is the most commonly misused chart type. You should only use a pie chart when you need to show relative proportions or percentages of information. The eye cannot quickly discern pie wedges and cannot compare one pie to another. If the viewer has to work too hard, key points from your data will be missed.
When to use:
- Showing proportions
Make it shine:
- Limit to five or less wedges. If you have more than five proportions to communicate, consider a bar chart. It becomes too hard to meaningfully interpret the pie pieces when the number of wedges gets too high.
- Overlay pies on maps. Pies can be an interesting way to highlight geographical trends in your data. If you choose to use this technique, use pies with only a couple of wedges to keep it easy to understand.
- Showing percentage of member vs. non-members, types of certifications
Figure 3: Pie used with a map. This is a visualization which combines a map and a pie to show purchase spending and membership status. Here we can easily see that Georgia has a large number of sales and more members than non-members. It is difficult to see the distribution of members and non-members in Mississippi, but clear that their overall sales volume is very low. Using a combination of charts like this does not allow you to know exact numbers or do exact ranking, but it does show relative proportions geographically.
In a future blog, we’ll review more specialized chart types that, while less common, can be extremely powerful in conveying the story your data is telling. Examples include, a map, scatter plot, histogram, text table, heat map, highlight table, bubble chart and treemap.