I feel like I should start this post by explaining that I'm not actually too sure of the name of these charts. Are they sun bursts? Are they spider web charts? Are they a niche form of radar chart? Honestly, I don't think they're any of these options but they're a fun chart choice to give a broad overview of some data without getting too deep into the nitty gritty. To explain how I got here without even knowing the name of the chart, I'll walk through my thought process.
In the dashboard captured in the feature image, I had originally planned to explore the performance of heptathletes (originally inspired by a Preppin' Data challenge on the same data) and the significance of being a specialist vs a generalist using the sun burst adjacent chart as my vessel. I had seen a fellow member of my training cohort create this chart and use it to explore superstore sales by sub-category (find their dashboard here), and I thought it would be an excellent chart for comparing my heptathlete performance at a higher level without the need to show the exact points scored. I thought the small multiples style could also offer an excellent opportunity to recognise patterns amongst the competitors but I never actually got that far because I got caught up on how to illustrate my measure.
As a result, the final product actually explored how the results of each event could be represented using different calculations of the same measure and I think this chart style really highlights those distinctions. My left hand charts demonstrated the raw points scored across the seven events, the 2nd column represented the percentile the athlete was grouped into in each event, the 3rd column explores comparison to the event average points and the 4th column explores comparison to the event maximum points. Thus, all using the same measure of points but the charts are able to show comparative differences. For example, you can see that in two of the events, the person who finished 3rd overall actually finished in the bottom 12.5% of competitors (as shown in the 2nd chart in). However, they still achieved at least 75% of the maximum points scored in those events (as shown in the 4th chart along) which is likely how they still managed to clinch 3rd position.
So, that's how I got to this point. Many sun burst style charts in, I thought I would spread my wisdom to the masses because when I tried to find out how to do this, there wasn't really a clear way to go forward. I learnt from word of mouth (and maybe some assistance from Gemini) but my favourite way to truly understand how to create a chart is using a step-by-step guide so that's what I am providing to you. I couldn't find one online, perhaps because I don't really know what the chart is actually called, but hopefully this offers you something new to try out. I think my heptathlon exploration is a good way to use the chart style (you can see the whole dashboard here) but so that you could follow along with a simple dataset, I've mocked up some sun burst adjacent charts looking at sub-category sales broken down by category. In this demo, you should be aiming to create something similar to those below. We might have some fun with creating all three charts in one at the end of this too (colouring by category) so stay tuned for that. I've rambled for far too long at this point but now you know where this all began, let's get to actually creating the charts.
Step 1: Map undertones.
To enable the layering of the charts without the use of a dual axis (which would only allow for two layers like a donut chart - I know, American spelling but for some weird reason in my brain ring donuts, like a donut chart represents, are less doughy so they're donuts and aren't worthy of being doughnuts), you need to use map layers. Start by writing the MAKEPOINT() calculation below:
This calculation is the foundation of your chart. Drag it onto the canvas as many times as you'd like, adding as a mark layer each time. In my example, I've got 6 rings plus an extra mark layer for the inner circle, however in my heptathlon dashboard I had 8 rings so it really depends on what works best for your data. I probably wouldn't recommend going much higher than 10 rings unless this is going to be a large chart on your dashboard because differentiating the rings on a small chart will get tricky.
You need to change the mark type to a pie and add sub-category to detail on each layer (except the top most one which will be your inner circle). I have filtered to a specific category as you saw that I'd split the sun bursts into three but it depends what you'd like to do. The rest of this blog post until the end will be focussing on one sub-category so I'd probably recommend doing that to make the follow-along easier but the secondary option will be available in more condensed form if you feel confident enough to parse which parts cross over with the earlier steps. Anyway, once all of that is done you can remove the longitude and latitude from your rows/columns shelves.
Step 2: Intermediate formatting.
Once you've got one central circle in the middle of your sheet (kind of like the Japanese flag but in blue), I find it easier to format it so that I have an idea of what everything's going to look like. I work from the outside in which means starting at the bottommost layer and working my way up so that I can see the maximum size my chart can be. You could go the other way around if you need your inner circle to be a set size though to fit a label. If you are doing that way, just remember that the chart size will be different when you add it to your dashboard so don't make things fit perfectly in your worksheet because you'll likely be disappointed when you reach the dashboard. You can see the comparison of worksheet vs dashboard for my heptathlete labelling below to give you a better understanding of the size differences.
To format the chart, you'll need to amend the size of the pie and add coloured borders to create the spider web style pattern. I like to add colours to distinguish the rings as well.
Step 3: Layer calculations.
To colour each layer individually, you need a calculation for each layer. Working from the outside in because the lowest numbered 'central points' are the outside ones, start with your maximum value that you are happy for your sales to exceed. Because I had 6 layers, I looked at the sales across the different sub-categories and felt the best compromise was each layer representing $50000. Thus, the maximum value for the outermost segment to exceed needed to be $300000. Copy out the calculation below for layer 1 and replicate it for the other 5 layers, stepping the value to exceed down by 50000 each layer.
Then drag these calculations onto colour on the appropriate central point marks card. Stagger the colouring as you see fit by editing the individual colour legends.
Step 4: Tooltips.
To ensure the tooltips are accurate for the different segments, you need another set of calculations for each individual layer. I think you could probably hard code it so that it specifies for each sub-category the maximum value that they exceed but I liked how each segment showed the different values. For example, as below, phone sales exceeded $300000 but if you focus on layer 4, you can see that it exceeded $150000 at this particular level.
It's quite a simple IF statement to do, just time-consuming as you'll likely find! Copy out the statement below for calculation 1, then edit it for the correct value and the correct layer activation (e.g. layer activation 2 for tooltip 2) for the rest of your tooltip calculations.
Step 5: Final formatting.
Add your tooltip calculations to each layer and format as you wish.
I sorted my chart by sales by dragging the sub-category on detail to the top of each marks card and then adding the sort order.
I also added a label to the outside layer so that the segments were easily identifiable.
Step 6: The other sub-categories.
To replicate the chart with the other sub-categories, I duplicated the sheet and then had to duplicate the layer calculations to be able to create new colour legends, replacing these new colour legend calculations with the old ones. You can see below that nothing else has changed (i.e. you can use the same layer calculation just with a different name and the tooltip calculation doesn't need to be duplicated as you're still using the same values) but it's enabled different categories to be different colours in the new charts.
The key thing here is that if you're trying to compare across categories, it's important to use the same value for each segment. Even though office supplies made fewer sales, it would be misleading if you changed the segment value (e.g. from $50000 for tech sub-categories to $20000 for office supplies sub categories) as a naive viewer would naturally assume that if the maximum sub-category per chart filled the outermost segment (phones vs storage in this example), then they made a similar amount in sales. Of course, that would not be correct so stick to the same segment values across your charts.
Step 7: One chart. Three categories.
If you've gone through all these steps but you decide you'd actually rather have all the sub-categories in one chart then that's okay - just remove the category filter. But you want to maintain the differentiation in colour between the different categories, I hear you say? Well that requires some more calculations! Or some editing of the original calculations, whichever you fancy really. I prefer to start afresh so I have both calculations available to me if I so choose.
To get the colours to be unique to each category, you need an if statement in your layer calculation. Well, an extra if statement. You also need to convert your category to an attribute as you're aggregating in the outer IF statement (summing your sales) so that you return only one value for each sub-category grouping. I'll be honest, attribute is a little bit of a mysterious aggregation to most everyone I speak to so just trust the process if you're still a little fuzzy on the why of it all.
Add your calculation to your central point layer and duplicate it for the rest of the layers, editing it so that the value the sum of sales needs to be greater than is appropriate for the layer. Edit the colours in the legends to make it pretty and you're almost there. You just need to sort the tooltip calculations for each layer so that everything works seamlessly.
And there you have it - a fun new chart to use in your next dashboard. Remember, this chart is best at demonstrating patterns within the data rather than the exact differences between values themselves so use it with caution. Still, it's something different and somewhat impressive that you've learnt so also use it with pride! That's everything from me for now. Until next time, happy charting!