I like bubble charts – even though they get a lot flack for being difficult to understand. These charts can pack a lot of data into a few variables. I agree the charts can be difficult for layperson to understand but it doesn’t seem true for the geo bubble maps. Possibly it’s because the user sees the map and understands its related to location. Here’s 3 tips for using a geo bubble map with your location data.
Using a Geo Bubble Maps to Combine Ideas
In the previous post, we created a geo-regional map to show the average damage cost from F5 tornadoes for each state. One issue with the method was that users had to hover over each state to see how many storm events were associated with each event. If a user wants details, it is a little awkward. A geo bubble map resolves this issue using bubbles.
A geo bubble plot places a bubble on the geographic location and allows you to control two aspects of the bubble – its size and color. In the following example, the bubble size is the event count (meaning number of tornadoes) while the color shows the estimated property damages (shown with the scale). Now it is more apparent that Kansas endured similar number of events as Alabama but the price tag was a little larger. However it also shows that Ohio had a similar cost but less events.
Tip 1: Add a Hierarchy to Allow More Data to Display
In the above example, I started out by showing the cost per ZIP code. It was a lot of data! So many that SAS Visual Analytics warned me “No data appears, because there is too much data to display.” The recommended fix is to filter the data, so I think a hierarchy might be a better way to go. .
The following figure shows how I used the Event state and custom data item Tornado F/EF Scale item (see Get to the point with a Geo Coordinate Map) to create the hierarchy. For step 4, you want to put your hierarchy in a large to small grouping. For instance if you were build a date hierarchy you would use Year>Quarter>Month>Day>Hour. Otherwise the hierarchy wouldn’t make much sense.
Now when the user clicks on the bubble state – it will then show all the tornado touchdown points. This is how Kansas looks. It does reveal a few areas where there was some significant cost involved. The gray bubbles mean that there was no associated damage with the touchdown.
Honestly – even though its helpful to see location – it’s a hard way to look at the data. It would be easier to see this breakout in a different method. Another idea would be to group the ZIP codes into their county so the data is not so widespread. My point here was just to show you how to create hierarchy and apply it. [Check Part 1 Is Location just a Distracting Character in Your Dataviz for an overview of the geospatial objects and when to use geographic data.]
Tip 2: Consider the Bubble Color Choice
By default, the geo bubble map uses a gradient scale of red to blue. This scale is usually good because you are showing performance as it increases. So red means bad and blue means good. However in this case – our data shows the property damage – technically any property damage is bad. We are not measuring how well the tornado was at damaging property (Sorry Tornado Gods – you will not get a raise for causing more damage!)
You can control the colors on the Style pane. I like to use a single color and just select the variations myself. I don’t think blue offers enough contrast – but I do like it.
In the end – I chose teal. It’s a nice compliment to the blue but offers some contrast.
Tip 3: Add Bubble-mation to Move Through Time
Ok – now for some real fun. We have shown the user the tornado location, the count, and compared the damages across the states. What if we plot the years to show how infrequent these events are? Right now the user can look at the chart and it may appear that each year a massive tornado makes a path through the US. However, the data is spanned across nearly 70 years. In reality – 1974 had the most damaging tornadoes.
Just by adding a date data item to the Roles – SAS Visual Analytics automagically animates the map. I used Event Year as the Animation value and boom – look what it does. Now the user gets a better idea how infrequent these events are as the animation moves through time. Here’s an example of 1968 – 1976:
This concludes my 4-part series about geospatial objects: Part 1 Is Location just a Distracting Character in Your Dataviz, Part 2 – Get to the point with a Geo Coordinate Map and Part 3 Use Geo Regional Maps for Overall Comparisons.
Never miss a BI Notes post!
Click here for free subscription. Once you subscribe you'll be asked to confirm your subscription through your email account. You email address is kept private and you can unsubscribe anytime.