Heatmaps and highlight tables use color to help visualize data displayed as a text table (crosstab or tabular view chart).
Highlight tables enhance text tables while keeping their form. They encode ranges of measure values with the preattentive attribute of color; from lowest to highest. These tables can display either continuous colors using sequential or diverging palettes. They can also use a stepped array of colors. A diverging color palette utilizes different colors to highlight crossing a meaningful threshold. An example would be going from positive to negative values. A sequential color palette varies the intensity of a single color to highlight rank.
Heatmaps go a step further. They use both color and size as preattentive attributes. This allows you to encode two measures: one for color and another for size.
How To Read Heatmaps and Highlight Tables
Highlight tables display data in a text table. Using color, they speed up how you identify the most important numbers within a range of values. These tables also have rows and columns to depict different dimensions.
Heatmaps present patterns, trends, and relationships within the data. They use color and size from two on the same or different measures. For example, while using color to help identify changes to a measure in a heatmap, the creator might reinforce its importance by keying the same measure to size. To identify trends you’ll want to ask yourself: where is my point of interest in relation to the lightest, palest color on the heatmap? Is there an increase in the size of the shape used at a certain time or place in the map? The answers to these questions could indicate a relationship or trend forming within the data.
What Type of Analysis Do Heatmaps and Highlight Tables Support?
Highlight tables are text tables enhanced through the use of color to show high and low values. Suppose you have a financial spreadsheet that tracks revenue, loss, profit, net income, gross income, costs, and sales. You can use different colors on the measure value for each data point labeled. If you care about highlighting only declines, you can use one color for negative examples (ex. red). Positive outcomes can be a contrasting color to mark the difference (ex. black).
Heatmaps work best for presenting trends in dimensions that have more variables. This is because heat maps consist of one or more dimensions and one or two measures. For example, a hospital can have many patients come in and out at the same time. The hospital may want to know whether these patients are scheduled or are emergency visits? A heatmap can make that distinction with unique rows or columns for each type of appointment. They will then show hospital peak hours, and patient volume with color and size.
When and How to Use Highlight Tables for Visual Analysis
Use a highlight table to indicate salient values on a text table using color. Viewers will be able to quickly distinguish the higher measures from the lower measures within seconds of looking at the view.
Bad highlight tables will have:
- Multiple colors that confuse the viewer
- Measures that are too similar overall, thus making it difficult to see where the higher and lower quantities lie.
- If they show more than one measure and have the same colors for both measures.
- Divergent color palette when all values are positive or negative
- This should only be used when your data has a meaningful threshold that is being crossed
Great Examples of Highlight Tables
This highlight table looks at the number of patients who are admitted in a hospital, and their respective wait times. The darker the color, the more patients there are in the waiting room at that time.
- The data is spread evenly. Viewers can see when more patients are waiting and when the waiting room might be empty.
- Elective and emergency wait times are split into two tables, but use the same color.
Ineffective examples of Highlight Tables and Alternatives
Poor Example
This highlight table looks at sales, profits and orders per month and year. However, this table is poorly designed and can leave the reader confused.
- One table uses a similar color to the table that looks at the number of orders. It is impossible to tell the difference between the low profit color and the number of ordered colors. Profit and sales also use similar colors that bleed into each other.
- Orders, profits, and sales blend in with each other.
Better Alternative
A better alternative for this would be a bar chart that looks at all the sales, profits, and orders per month and year. The numbers are more evident in the bar chart than they were in the highlight table.
Furthermore, the categories don’t bleed into each other. Highlight tables shouldn’t be used to measure more than one category.
When and How to Use Heatmaps for Visual Analysis
Use a heatmap to show a trend, pattern, or relationship in a set of data that fits well on a table.
Bad heatmaps will have:
- Too many similar sizes across shapes
- The size and the color represent two different measures
- Not a strong enough distinction between big and small numbers
- Unreadable text
Great Examples of Heatmaps
This heatmap looks at sales per region and subcategory. The viewer can see clearly that some subcategories, like chairs, sell very well in all regions, while others, such as art, don’t sell very well anywhere.
- The color theme remains the same throughout
- The circles aren’t all similar sizes
- Only four regional markets are measured, making comparison easy for the viewer.
Ineffective examples of Heatmaps and Alternatives
Poor Example
This heatmap splits those four regions used previously into four years per region, making it difficult to compare the same Sub-Category across Regions for the same year. The heatmap also relies on legends to decipher what mark color and size means which can make it difficult to understand.
Better Alternative
Because a heatmap is still the best solution for this much information, we will keep the chart type but will reorganize fields to make it easy to compare subcategories for each Year/Region combination. We’ll also change the shape to one that is easier for our eyes to compare across space (square). Last, we will put descriptors in the title making it easier to identify what the encoding of size and color mean in the view. Here is the proposed alternative: