Interpreting a Heatmap

Learn how to read and analyze the data that is synthesized in the heatmap visualization.

Victoria Kinzig avatar
Written by Victoria Kinzig
Updated over a week ago

In your reports, you will see a "heatmap" visual that synthesizes your quantitative information in order to highlight trends in the data. This article will help you understand how to interpret the information in your heatmaps.

1. Categories and Individual Topics:

In the left column of a heatmap, you'll find the individual topics that have been measured. These individual topics are grouped together for easy viewing and aggregation under "Categories."

Each individual topic corresponds with a particular scaled item on a survey, observation form, or other instrument.

2. Averages: 

When you land on your heatmap, each numerical cell to the right of a given topic shows you the average response, based on the scale that was used (usually a 3-point, 4-point, 5-point or 10-point scale). To see exactly which scale was used and how the average is calculated, hover your mouse on any cell. 

How are "individual topic" averages calculated? For individual topics, each unique response exactly matches with a specific point value (e.g. 1, 2, 3, 4, 5), and the numerical cell simply shows the average of all of the response values for that item. The color that displays for that average is automatically generated based on the ranges shown when you hover on a cell. Please note that, even though you see average ranges corresponding with each color, each individual response matches exactly with a whole point value.   

How are "category" averages calculated? "Categories" are not typically measured on their own. They simply function as visual aggregations of the individual topics. The number you see on a category cell is calculated as follows: First, we calculate the overall category average for each unique response (behind the scenes), based on how that person responded for each of the individual topics. This provides an estimation of "how is one person or one response rating overall with these related topics?" Then, we simply average all of these unique category averages to generate the number that shows on the parent topic cell. This gives us an aggregation of "How are several people or responses rating, overall, for the topics in this category?

3. Groupings:

Aside from the "All" column, you'll notice that the columns on your heatmap provide a way to segment the responses (e.g., by school, grade level, time of year, etc.). To switch the way the data is segmented, click on the "GROUP BY" dropdown above the heatmap, and select the attribute by which you'd like to separate responses.

When you look at your averages, they are representing the average of only the responses that correspond with the particular group or segment of the data you've separated.

4. Percentages: 

Clicking the "Percentages" view (a button just above the heatmap) helps to create "data soundbites" that are easier to articulate to others.  Instead of saying, for example, "The average in Standards Alignment is 4.1," you can say "81% of teachers have been identified as effective or highly effective in Standards Alignment."

How are the percentages calculated? The percentage that displays in each cell is calculated based on the proportion of responses that are on the 'high end' of a given scale. For example, with 5-point scales, you'll see the percentage of responses to an individual topic that rated at the 4 or 5 level. With 4-point scales, you'll see the percentage of responses at the 3 or 4 level. With 3-point scales, you'll see the percentage of responses at the top tier (3) level.

How are the colors generated on the percentages view? The color that shows is automatically generated based on which of the following tiers the percentage falls into: 

*Note about averages or percentages that fall "on the edge" between two color categories. 

  • As scales build from 0.0-5.0 or 0%-100%, each new tier will take effect when the datapoint is greater than (not less than or equal to) the ending value of the previous tier. For example, if the percentage seeing success is actually 20.0001%, this will fall into the 20-40% range, but will display as 20.0%. If it is 19.9999, it will fall into the 0-20% range, and still display as 20.0%. 

Did this answer your question?