I love heat maps – they’re a great way to show certain types of data in a really visual way. While typically thought of as being used to show the density of something, we can be quite creative with what density means. I look at a couple of helpful visualisations I’ve used in this post!
To show outliers or points of interest
When data is displayed in tables or individual charts it can be challenging to find patterns or outliers across them. In some contexts, heat maps can visually display outliers in a really nice way. One example of this was when we looked at the response times for a service which received requests from multiple sources. By placing buckets of response times across the x-axis and by displaying each of the sources up the y-axis, we produced a chart showing the density of response times for each source next to each other.
We could see that one source in particular (source 3) was receiving responses more slowly than the rest. This immediately gave us an area of focus to understand more deeply, more quickly than if we had built up several individual charts and attempted to correlate between them.
To compare state in a visual way
It can be tricky to explain changes in a context that might be subjective or hard to conceptualise. A recent example from my experience was communicating how ‘much’ technical risk there was in my portfolio of products. I had a one-dimensional rating which was useful for showing that we were resolving the greatest risks, but this didn’t move when we resolved some of the smaller, less impactful risks. This lead to work to visual risk in a heat map based on an x-axis of “likelihood” and a y-axis of “impact”, each of which were defined strictly beforehand. This lead to a chart which shows the density of risks across the entire risk profile.
Not only does this allow us to show we are making progress, but we can compare with an ideal future state and set goals around it. See the example below where we have decided that our goal should be to have a risk profile where impact reduces with likelihood. We can put these side by side to demonstrate to others how close we are to achieving that state.
Heat maps can be used in lots of other ways – I’ve seen them used to show geographical data, movement data across a space, and to show correlations between data sets that might not have been considered before!
Have you seen any great ways heat maps have been used?