When most people speak of data visualizations, they usually mean charts, but charting is only one way of making information visual. An organization chart, a picture, an infographic, or a quotation can have visual impact, depending on how you present them.
What makes a data visualization useful is how we use the design components to convey an idea, using the audience’s expectations as a guide. A effective data display draws attention to a main idea and motivates us to understand its meaning.
Visualization Tools in SumTotal Advanced Reporting
SumTotal Advanced Reporting gives us three groups of display tools for visualizations: tables, crosstabs, and charts. You create these displays from Ad Hoc Views.
You can use a delivered domain as a starting point. We recommend you copy the domain to the Domains folder (or a subfolder) in your Custom folder. We also recommend using Topics within your custom domains to help organize your reports in a way your users can find them.
After you have selected the Domain or Topic, add the Fields, Groups, and Measures you need to include in your Ad Hoc View. If you need to rename fields or create calculated fields, this is the place to do it.
When you have all your Fields, Groups, and Measures in your data set labeled, and arranged in the order you want, select the visualization type from the Table Type drop-down list. Let’s examine each type.
A table is a formatted list of detailed data that a user can group, sort, filter, and summarize. Tables are useful when a user needs detailed information about many items, such as a group of employees or learning activities. They are also useful when a user wants to receive data in a worksheet for further analysis and formatting.
One way you can make tables more useful is to create conditional formatting to highlight values and draw attention. Suppose a manager needs to know which employees have certifications that will expire within 90 days. You can create a conditional formatting rule so the Days until expiration Field displays a bright color if the value is less than 90.
However, if knowing which employees have not completed a course or task is the only purpose for the table, it may be more useful to filter it, so it only shows those items where the cell value is less than 90. If the report has other uses, conditional formatting might be best. Test a prototype with your managers to see how they use the information.
A Crosstab is a pivot report, so called because users can change how it summarizes data by pivoting rows to columns or groups, columns to groups or rows, and groups to rows or columns. You can change fields to measures to count them or measures to fields so you can group them.
Users can change how data is grouped and summarized and can drill down from a total or subtotal to a table of detailed data. This enables you to present a summary table that would show which items might need action, then allow the user to drill down to the detail to understand what action might be appropriate.
Using Crosstabs, reports analysts can quickly create summary reports that would otherwise take many times longer to develop. Crosstabs are also useful for savvy business users who want to examine data in different ways to better understand its meaning.
A crosstab is rarely a complete product when you create it. You will get much better results and have much happier users if you explore its formatting and summary features. Take advantage also of calculated fields to create a display that will aid better decision-making. The easier you make crosstabs to manipulate, the more users will adopt them. Don’t leave the heavy lifting to your end-users.
Don’t expect users unfamiliar with crosstabs to adopt them right away. Understanding how to use them will take training and practice. Start with a group of savvy managers to test prototypes and let them be part of your development team. Let them create the winds of change that will develop your managers into better data users.
Good charts are an efficient and powerful way to help people grasp information. They can create understanding across business functions, academic disciplines, nations, and cultures, without the need for a common language. Charts surround us every day in advertising, fitness tracking, utility bills, news articles, and social media. We humans use them to understand, explain, persuade, motivate, entertain, sensationalize, mislead, and outright lie.
What is a Good Chart?
Many considerations go into making a well-constructed chart: scale, orientation, typography, color, and contrast to name a few. Modern charting software will automatically create a nicely formatted display from the data you select, and most will prevent you from making common formatting errors.
But in the context of learning and learning management, a well-formatted chart is not a good one if it doesn’t deliver insight into actionable information. To do that, it must draw attention to the main idea and convey immediate understanding.
A simple way to test a visualization is to present it to the person who will use the information without any explanation. If your tester immediately begins drawing the correct conclusions from the display, you have succeeded. If they must ask what it means, you have work to do.
Principles for Good Chart Design
You can find hundreds of resources that tell you “do this, don’t do that,” but few explain the principles and the science behind them. Our discussion here is based on the five principles in Scott Berinato’s Good Charts: The HBR Guide to Making Smarter, More Persuasive Data Visualizations. We haven’t found a guide that mirrors our lessons of experience better than this one.
The Jaspersoft engine in SumTotal will attempt to follow the rules of good construction, but understanding these few principles will get you started on the path to good charts your people will understand.
The order we view data visualizations depends on shape, colors, texture, size, orientation, and a host of other factors, including how expert we are at reading charts. Don’t expect your title or subtitle to explain the chart. It might be the last thing your audience sees.
- A good chart draws our attention to a single concept: a difference, change, intersection, peak, similarity, or a dominant color. The two charts below demonstrate this concept. The first thing most people will first see the dip in the chart on the left and the rising trend in the one on the right. If something stands out in your chart in a way that grabs attention, it must be the main idea you are trying to communicate.
- The more data points you have, the more singular the idea it conveys. In other words, too many data points tend to merge into a single meaning. If you wanted to show how a training program impacted the performance of 12 service teams, displaying them together will obscure the results. A manager seeing this chart would likely conclude that nothing changed. Better to show 12 charts than one so your audience can gain meaning from each.
- We grasp visual cues thousands of times faster than verbal or written information, and immediately try to make sense of it, and often before we are consciously aware. We even try to make sense of things that look wrong. Look at this spurious example of the number of certifications completed in an organization. The number of certifications completed is meaningless, but your mind will still try to understand why Operations has so many more completions than other functions.
This chart helps you make the correct assumption and may make you want to explore why Sales and Support are lagging.
- We agree on many common assumptions: north is up, south is down; a line moving up and to the right is good; time goes left to right. Green is good and red is bad, except when red is hot and blue is cool.
Much has been written about colors, but modern charting software like Advanced Reporting uses good color management by default. What it doesn’t do yet is manage semantically resonant colors. For example, money is green, anger is red, oceans are blue, and mountains are green or brown, unless they are snow-covered. We may see tools in the near future that help us with matching colors to concepts soon, but in the meantime, we need to take care.Take a moment to look at a U.S. national weather map. Imagine your confusion is the warm regions of the South were blue, and the cold areas of the North were red. And what if you were looking at a map of the earth and the oceans were red or yellow?
Now, try to make sense of this chart:
Confusing, isn’t it? You don’t know at first glance which is the “good” number.
If you ignore conventions and metaphors, you can expect your users to be confused. SumTotal Advanced Reporting will attempt to obey our shared assumptions, but it won’t hurt to check with your users to see if anything goes against the grain.
The Last Word
The best advice we can give you is to make your end users, the people who will use the data, part of your development team. Listen to their needs and ask them for feedback with a relentless focus on their satisfaction. Remember also that your users will often not know what they need until they see it. Feel free to experiment.
1. Berinato, Scott. Good Charts: The HBR Guide to Making Smarter, More Persuasive Data Visualizations (Kindle Location 769). Harvard Business Review Press. Kindle Edition. 2016
3. Lin, Sharon, et. al. Selecting Semantically-Resonant Colors for Data Visualization. Computer Graphics Forum (Proc. EuroVis), 2013.
Pixentia is a full-service technology company dedicated to helping clients solve business problems, improve the capability of their people, and achieve better results.