Sign in or start a free trial to avail of this feature.
4. Tables, Heat Maps and Tree Maps
Heat maps and tree maps enable you to color-code larger data tables. Learn how and when to use these visualizations in this lesson.
To explore more Kubicle data literacy subjects, please refer to our full library.
- An alternative to text tables that color-code cells by the value within the cell
- Use when your manager prefers numbers to visualizations
- Limited to visualizing a single measure effectively
- Can visualize two measures by color and by size
- Can be combined with numbers although can make the map harder to read
- Great at visualizing changes over monthly or yearly time periods
- Display the size of a measure by the area of a rectangle
- Can be manipulated to display different hierarchy levels within a dataset
- Cannot be used to display changes in a measure over time
Sometimes you will have a manager or a client who will prefer to look at numbers rather than visualizations despite your best efforts to convince her otherwise. In this scenario, you may end up with a visualization like this, which simply uses the text tables option in the show me dropdown.
All is not lost, however. It's still possible to turn this chart into something that can be more visually engaging using highlighted tables. Highlighted tables are similar to text tables with the only difference being that you can color code the background of each cell. When I hit highlighted tables and adjust the entire view, I can see that my chart is now easier to read and I can pick out some good and bad months for each of the fee earners. I can make this easier to read by changing the default color scheme, so I'll hit the dropdown and again, hit the palette dropdown.
Here I'll select red and green diverging. Always tends to work pretty well. Then I'll press OK. As you can see, this chart is now much easier to read, where I can pick out the months in which fee earners are performing well, the drop-off in November, and also the individual fee earners who are having very good months, which can be hard to find without this color formatting.
In some respects, this chart can be improved if we remove the values.
So I simply drag the label outside the marks card, and again, some things are now easier to see. For example, the fee earner Salmon's performance drop-off from October to November, which seems pretty stark.
However, with the numerically driven manager or client, you probably won't get away with removing the numbers altogether. So I'll simply add them back with Control + Zed. In this example, we have one measure showing which is the sum of value. If we want to include multiple measures in a similar type chart, we need to move to a heat map.
To do so, highlight the show me button and select heatmap.
Again, I'll just change the view to entire view.
What we can see in this chart is that what used to be color-coded from red to green is now coded by size from large to small. In this new chart, what I'd like to have is value staying from red to green but the number of minutes from large to small.
So let's drag the sum of value outside our marks card. Let's put minutes for size, let's put value for color. This gives us even more insights than we saw in the highlighted tables.
What we can see is that although Jalene Yoon has the best month of all in October, she doesn't actually work the longest. In fact, Bennett Healy and Karlie Grady work longer hours than Jalene but are simply not charging as much.
Clover Calwell in October also shows this example where you don't have a lot of minutes charged to the client but you do see quite good fees generated.
If your boss likes this chart but still wants to see some values in the visualization, then it's very easy to do so. So let's take the values and apply them to label and just make sure that all of my sales people are selected. To my mind, this actually makes the visualization less effective but it may be a compromise that you have to undertake. If you like these types of visualizations where we combined color and size, you might also want to check out tree maps, which work when we want to include more than one dimension and we only have one measure. For example, value, which represents fees generated. Let's remove this daytime value from the columns.
I'll also remove minutes.
When I go to show me, I'll select the tree map option.
What the tree map option shows me in the shape of a rectangle is the value created by each of my fee earners over the time period and the size of the individual rectangles relate to the value.
One nice aspect of tree maps is that you can create an additional dimension that drills down further into the detail. So for example, if I'd like to split this rectangle for Jalene Yoon into client, I simply select client and drag onto label.
And now I have Jalene Yoon split by all of her individual clients.
It also shows me the largest clients for each individual salesperson and visually tells me how much they contribute to their total sales. So for Karlie Grady, we can see that she has two clients that contribute the vast majority of her fees generated.
The color coding on this chart is still based on the sum of value. To make it easier to read, I'd like to switch this to fee earner.
So I'll take the sum of value away from my marks card and include fee earner.
As you can see, this makes my chart a lot easier to read. I'd encourage you to experiment with tree maps, heat maps, and highlighted tables, particularly when you wanna show a lot of data points on a screen and where color and size can work very well together.