A fairly common problem in industry is how to visually show multiple variables across multiple dimensions. For example, a business might want to visualize how different sales regions compare for several different product or product lines. Or, maybe a nonprofit would like to show how countries compare on rankings of several health or social indicators. Another use case might be to show how certain products compare across multiple features. Or, perhaps you’d like to show how how a set of cities rank across multiple categories like I did to show the 2016 Clean Tech Leadership Index ranking for U.S. cities – showing how 20 cities rank across 4 rankings and a composite as shown below.
Parallel Coordinates are a good way to represent this kind of information, and, thanks to tools like RAW Graphs, are pretty quick and easy to create once you know how. The first thing you’ll need to do is get your data into the right format. Using the example of the clean tech city index from the image above, the format I needed is shown below. The five vertical axes that will be represented in the graph are the data in columns C – G. Column B in the spreadsheet is not used in this example.
Once you have the data in this format, you just need to go to RAW Graphs, paste in your data, then select the Parallel Coordinates graph. From there, you just need to map the dimensions by dragging the corresponding column heading names to the dimensions and in this example dragging the Cities header over to the Color area as shown below.
This will result in the basic graph shown below.
Some Tricks to Improve the Parallel Coordinates Chart
You’ll notice that this is pretty basic. There’s a few things to point out that you might want to improve upon and a few tricks and techniques to make this chart better. For example, the vertical axes are different scales. This is fine in a lot of situations when you’re visualizing very different dimensions. But in this case, each index’s dimensions are on a scale of 0 to 100 so it kind of makes sense to normalize them. You’ll also notice that the chart doesn’t include labels for each of the lines so that you can better identify which line corresponds to which city.
The trick to get your axes to all be on the same scale of 0 to 100 is to add a couple of rows to your dataset. Notice in the image below that I’ve added two new rows to the dataset; Max and Min, each with a corresponding value of 100 and 0 in each column.
Now if you paste this data into RAW Graphs and repeat the steps to generate the Parallel Coordinates, you’ll get something like the following. Note, this little trick only makes sense if you plan to download the image as an SVG file and edit it in Adobe Illustrator.
Now, to label the lines in the chart, you’ll want to export the chart as an SVG file and open it up in Adobe Illustrator. Unfortunately, the current version of Parallel Coordinates in RAW Graphs doesn’t generate the labels like it does for some of the other chart types. So, you need to added them manually in Illustrator by adding text labels and positioning those labels adjacent to the start and/or end point of each line. But how do you know which line corresponds to which city? Here’s a little trick. Go back in RAW Graphs where you can adjust the hex color for each item (screen shot below). Notice how they’re in alphabetical order?
Well if you look at your Layers Palette in Illustrator where the paths are, these will be ordered in reverse alphabetical order.
So, a handy thing to do is to go in and relabel each of those layers, in my example with the name of each city in the dataset, thereby making it much easier to associate the labels you’ll make with each corresponding path. Here’s how my Illustrator file looks after I’ve gone through and renamed the layer names, added text labels and made a few other aesthetic adjustments like modifying the stroke color and stroke width. You’ll also want to delete the paths representing the fake Max and Min values entered into the dataset to get the axes all normalized.
So next time you need to show several things rank across multiple indicators, whether in a blog post, press release, annual report, or white paper, consider using a parallel coordinates chart using RAW Graphs and Adobe Illustrator.