Join Bill Shander for an in-depth discussion in this video Rethinking some charts, part of Data Visualization for Data Analysts.
- So, in this movie I'm going to share a bunch of examples of charts and show you the originals, pretty much as I found them, and then walk through how I redesigned them applying the principles I've been talking about for this entire course. Now, you might think I'm beating a dead horse, but I have to say it again, the main point here is to simplify, to do less, to know what you really want to say. That's what it's all about. So, here we have a chart. And the scenario here is we have, let's say, data from two companies. One's a manufacturer and one is a retailer and they've agreed to sell the products at a certain price.
And the retailer is actually just trying to make the point that when we are under contract to sell the product at a certain price, we adhered to those rules, to certain amount. And, even when we weren't under contract to do so, we actually even adhered at a higher amount. So, they were doing a good job even though they weren't being forced to. Essentially, that's the idea. So, this is the original chart. And I'm gonna show you the redesign. The redesign is just much simpler and takes out a lot of visual noise. And, also it follows all the other things that I always say to do. So, for instance, if I just go back and forth.
My axes are now a gray instead of black. You'll notice I also took almost all of the numbers out of the axes because who cares about where 20% and 40% and 60% are. The point of this one is the trend line sort of overall, the shape of that line and then relevant position of those lines. You'll notice I also took on the original axis on the left-hand side, the y-axis, Cumulative percentage of SKUs, on the left-hand side there, I have to turn my head to read it. That's not very helpful. So, I just moved it up to the top and sort of made it part of the number, 100% of SKUs (cumulative).
And I just put a zero percent at the bottom, etcetera. You'll also notice I took out those two big arrows. They're not really helping me understand. I understand that if higher adherence is in the upper left, and lower's in the bottom right, that that's a directional thing. So I made them nice and bold, but not too bold. They're still gray. But, you know, all caps, really can draw the eye, but not too, too much. And just stuck them in the corners. Got rid of those arrows. You'll also notice, I got rid of that grid. If I don't need those hash marks and those percentages on my axes, do I really need the grid to point out where those lines cross each other somewhere in the middle of that chart? I don't think so.
And finally, in this example, you'll notice that we have a key at the bottom which, as I've said before, when you can avoid it, get the key out of the bottom. You don't want it off to the side where someone has to look back and forth, back and forth to figure out what they're looking at. Just by putting the key in line. Right, Outside Agreement is in the same color as the line it represents. It's right next to it. It's really easy to follow what's going on here. So, here's another example. Here I have Manufacturing Cost Advantage. The title's pretty obvious. The point here is that I have different brands of some kind of machine. And the idea is that the more pieces you produce, using this machine, the cheaper it is.
Makes sense. You'll notice that we have a logarithmic scale. In my redesign I did not touch that. I'm assuming that that's an important piece of this. I didn't want to redo the scale. So, here's my redesign. Really quick. Kept that scale but took out all the hash marks. The hash marks aren't visually helpful. They're not helpful from any standpoint. You'll notice I also took some liberties. And so I made the assumption that, if I go back to the original here, here they listed every single brand. And they had these colors as well as those little hash marks, the little shapes, so I could easily follow which line I'm looking at.
Well, you know what, I can't easily follow it. It's visually confusing. So, I made sort of a leap of faith here and said, you know what, probably they don't need to compare twelve different brands. I'm guessing really it's probably a couple of categories of brands that could be singled out, so I used contrasting grays to show those. And then really focus on the two brands that actually matter. Those are the two companies that are the sort of point. What do I really want to say in this graph? So, two bright, colorized, thicker lines to draw the eye, and everything else is simplified just like we've done before. Here's this example which we've seen in earlier movies in this course.
This thing's a hot mess. I hate this chart, but I love it. I love to hate it. There's so many colors. There's so many segmentations. There's different types of column charts. There's call-outs. There's multiple headlines. I mean, everything that can go wrong did go wrong on this chart. So my redesign, very, very simple. I took it as faith that the headline, so if I go back to the original, the top line, I'm assuming that's the actual headline. So, I made that huge, "The Outlook for Revenue Growth is Dim." So, what does that mean? What is the actual story? What is that headline trying to tell me? I'm guessing. This is about 2014.
This must have been 2014 data looking forward to 2015. I can tell by looking at the data. So, I highlighted 2014 to sort of show that that's this year. It's worth drawing attention to that. And that call-out that was fairly de-emphasized that this year went down two percent. Let's make that pretty big and bold but then, of course, the point is about 2015. A base assumption, a bear market assumption, and a bull market assumption. So, colorize those three columns. Draw attention to them. Put the percentage right above it. They don't need to be sit in circles and really draw the eye in some special way.
And then all the rest of it is grayed out. This is the point, so let's focus on the point of the story. Here's another example. This is a government chart from a government report. So, what are we looking at here? Mean GDP Growth. And, I don't have a problem with the chart form, in this case. I don't even have a problem with most of the components of this chart, even that background gray rectangle to indicate a recession period. What I don't like about this chart are those big, honking, thick diamond shapes and the thick lines. And the fact that everything is sort of overwhelming.
Plus labeling every single point on the chart. So, one approach to the redesign. This uses a lot of color. Maybe it's too much blue, but you know taking everything and just sort of simplifying the shapes. And shrinking down those average middle points, that mean GDP growth number, those little dots in the middle. This is just visually simplified. I also added an overall mean as a little dotted line across. So, it's not like I took everything away. You can do a lot even if you visually simplify a chart. Here's another approach to the same chart. Maybe the idea here was to actually call out a couple numbers, fine.
I'm not opposed to labeling things. I've said that before. Just label what needs to be labeled. Don't label everything. You'll notice in this example, I've also removed that blue line around the recession period and made that axes gray instead of blue. And look at how much simpler that is. So I do think that this is objectively not as strong as this one. It's just easier and simpler to really draw the eye to what you want to draw the eye to. In this case, I do have a key in the upper right-hand corner. I really couldn't find a way to put a key in line. Sometimes that's okay. You got to do what you've got to do.
And a third approach. Okay, you've convinced me. You have to label everything. This is a government report. Okay, label everything. But there's a way to do it that is visually appealing that is not too overwhelming. And another example, here we have a chart. If you look at this closely, we have in line keys, right? So, Consumption-to-DPI ratio is in green and Total Wealth-to-DPI ratio is in pink. That's good. They did that in line. Unfortunately, what they also did was they tell us which axis to look at in that legend in line. And so it's really confusing. There are two axes.
There's a lot of text going on in the middle, here. It's visually confusing because of all those recession period gray background lines. So I took this one and simplified it. I took some liberties. First of all, those recession background shades, I dimmed way back. Maybe you need them, but let's not make them quite so distracting to the fore-front. But the main thing I did here is I split this into two charts. I don't want to have one axis on the left and one on the right and the labels to tell which is which. How about two different charts? I can see them on top of each other. I can still compare the shapes of these lines to each other. But the fact is that putting them on the same chart with different axes, it didn't make sense where they lived next to each other.
I couldn't tell one being high or one being lower actually was a valid point. But I can still compare the shapes of the lines which is still pretty interesting and useful here. And my last example. So, here we have a very complicated chart to look at and the explanation is in the text below the chart. If 5% of the adjustable market uses this machine, Platform A, and they each use that for 30% of the year, these manufacturers would produce 2.4 times the current output for this sub-category, 20% of the total market for this category.
So, the point here, I guess, is where that red line crosses that hash mark at 20% on the y-axis and 30% on the x-axis. So, I get it. It took me a long time to understand this chart. But, once I got it, and decided to redesign it, I came up with a couple different approaches. So, one of them is this one. So, you'll notice I've bolded the line for that 5% of addressable market. Okay, this is the line we're focusing on. And, I took out that big, honkin' orange arrow pointing at my text and connecting that line to my text.
You know, I don't need to do that, because using color, everything is gray except the things in orange that I'm drawing your eye to. I'm making it clear that the 20% and 30% and that point where they meet is that scenario. Again, which is in color, at the bottom, in orange. You'll also notice I moved the axes. I sort of took that right-hand axis and put it as parentheses within the left-hand axis. So, Market Share versus Output Growth. I sort of combined them even though they are still two values that I can look at at the same time. Now, in this case, I actually added to the grid.
Here we have a grid of lines across from the y-axis. Well, you know, the point is where both lines cross. So, I added a grid also for the x-axis values. But, you'll notice I grayed it way back. It's very translucent, very light colored gray so it's not as visually disturbing and distracting. I also took the in line key, you know the labels for Platform Penetration 10%, 7.5%, etcetera, and sort of pulled that outside of the box. Again, just to make it visually less distracting. Here's another approach to the same chart.
Here I took the grid out because I'm actually circling the four points I need you to look at. And it's still drawing attention to the main one, but sort of showing those other scenarios. What if Utilization is at 50%, what if it's at 20% for those different platforms? So, the point here, as you can see as I've redesigned some of these charts multiple times, is there may not be the right way. In fact, there is not the right way to visualize anything. It's always about options. So, again, bring me back to the beating that same drum. It's all about telling the story.
Telling the story you want to tell and finding the best way you can think of to bring your story to life in visual form.
- Why visual communications matter, and how they work
- Communicating via story
- Communicating with color
- Using legends and sources
- Sketching and wireframing
- Rethinking slides, charts, and diagrams
- Rethinking your templates and brand guidelines