Avoid Tables, Demand Visualizations!

Drew Skau
Written 2 years ago
in Design

The great thing about information visualization is it turns numerical information – which takes a high cognitive load to process – into visual information, which takes a very low cognitive load to process. Often, this process involves turning a table of numbers into a bar chart, line chart, scatter plot, or some other chart form.

Yet, some designers don’t bother turning tables into visualizations. Sometimes it’s because of a tight deadline, others: too much work, last-minute additions of data, or simply not knowing any better. Whatever the excuse, though, it’s not good enough. If your visualization contains a table, you can do better.

Let’s look at an example. Below is a table used to show salary ranges for a few professions in different cities in the US. The table has been colored based on the profession, but the colors really don’t convey anything that isn’t already shown on the column labels. (To help hide the identity of the creator, the profession titles have been cropped out.)

Spreadsheets are pretty good when you are working directly with numbers, but unfortunately, this is a static table. Worse, it is an image of a table, so copy and paste is not an option. All the wonderful functionality that comes with a spreadsheet format is missing. Even more unfortunate, this particular graphic uses data scraped from a search-based site, so in order to play with this data it had to be transcribed by hand (transcribed version here).

Could we show this information better than this table? Definitely.

One option is to do a series of bar charts. This version makes sense for people who want to look for the best place to practice their profession (the best-paid one, anyway). Comparing dollar amounts no longer requires thinking about which number is greater: now it is as simple as looking at bar sizes.

This method is very repetitive, though. The city names are labeled six different times. If we combine these charts into one chart, we can eliminate a lot of reading. Admittedly, this chart has a lot of information, and it would take a while to digest, but it does point out something interesting. For every city, except for Detroit, there is a predictable pay difference between each profession. (This is actually pretty suspicious, and calls the original graphic’s sources into question, but let’s ignore that for now.) Would that pattern ever have been visible in the table of numbers? Would you have ever seen the pattern get broken twice in Detroit?

This data set does have location information in the form of city names, so we could even use a map to represent the data. There is only one map here, showing just the 25th Percentile for Profession A, but a creative designer could easily use bar charts instead of circles, or come up with some other glyph to show multiple percentiles and professions at once. Now, a map is probably not the best way to show this specific information, but sometimes there may be regional patterns caused by climate or culture, and a map can help those become visible.

There are many more creative options for visualizing this data, all with varying degrees of success. For making the data more easily digestible though, almost all of these options would be better than a table of numbers.

If you are a designer, and your client pushes too hard with deadlines, or tosses data in at the last minute, do your best to push back! If you are a client, you should be demanding good visualizations from the designer you hire, don’t let them get away with a table. It is in everyone’s best interest to communicate data clearly and efficiently, and a table of numbers just doesn’t do the job.

Drew Skau is a PhD Computer Science Visualization student at UNCC, with an undergraduate degree in Architecture.