Heatmaps are two-dimensional representations of data that show magnitudes as colors. A color variation - may be based on hue or intensity - gives a clear indication to the reader of how the data measured is distributed or varies over space.
The result should be a data piece that doesn't require a lot of explanation. Typically, warm high-contrast colors are used to represent high-value data points, while cool colors represent low-value data points in heatmaps.
Most common Birth Dates (US)
Qlik Sense Feb 22 (day of the month/month)
The chart above is an excellent use case for a heatmap, it is simple and straight to the point. Births are fairly evenly distributed throughout the year, with a slight bias towards the summer months. The average number of births drops significantly on some days, such as 4th July, 24th-25th December, and January 1st.
There is also a very popular version control product out there that displays a punch-card-like activity chart like the one below that I think is an excellent implementation of a heatmap.
Not my profile
With the February 22 release, the Grid Chart will be your first choice when it comes to heatmaps. In addition to all the features that one expects in a Grid Chart, there is now a shortcut that lets us create a heatmap just like that.
Remember that this chart is especially useful if it is self-explanatory, so try to avoid complex combinations of dimensions. This chart is best suited to time-series-based dimensions.
This is fine, but it misses the very valuable ability to represent a 'third' dimension based on the relative sizes of the rectangles. Will this be available in future updates to the Grid Chart?
@cfountain72 Do you mean to add a 3rd dimensions to make the symbols pie charts (like the good ol' QlikView days) or a 2nd measure to make the size of the symbol different from the number that appears?
You can already do the 2nd thing by defining your measure using the dual function. The text parameter of the dual function appears as the label and the size of the square is based on the number parameter. The following chart has the measure Dual(Num(Sum(Sales)/100, '#,##0'), Rand()) so the size of the square has nothing to do with the actual sales.
Oh, and don't forget to format the measure as Measure Expression for this to work. That always gets me!
Sorry that I wasn't precise on my terminology. In a Heatmap (maybe better described as a Tree Map), you're not confined to an x/y grid, which allows the dimension (in this case, a hospital department) to be sized based on a measure, and colored by a different measure. In the example below, the size represents the volume, and the color represents a turnaround time. There is no need to have a second dimension as is required in the Grid Chart.