In cartography there has been almost a revival of heat maps in recent years. I should start by clarifying what I refer to by the term ‘heat map’.
Across different disciplines, heat maps mean different things from isotherms in meteorology to pixel matrix displays in graphics and computing. In cartography and GIS, the term is used to represent location weighting or densities. Caitlin Dempsey gives a very good description on GIS Lounge:
‘Heat mapping, from a geographic perspective, is a method of showing the geographic clustering of a phenomenon. Also known as hot spot mapping, heat maps show locations of higher densities of geographic entities.’
My research into heat maps and learning how to create them within a GIS has led me to the conclusion that heat maps can be a little more complex than just densities. They can take into account other factors such as proximities and weighting. Perhaps due to my academic background in Mathematics and the Sciences, I almost view them more as probabilities. Areas of likeliness of something occuring, for example an area where you are more likely to find a pharmacy that is open, an area where you are more likely to find pensioners resident, and so on. In any case, they are a valuable visual method of displaying geographical, location-based data or statistics.
The graphical matrix form of a heat map, a closely related cousin to the GIS form, which is like a table of statistics with each entry colour-coded rather than given a unique value, dates back to the late 19th century. Today, true cartographic forms involving geographic feature geometries are becoming common place, albeit with their critics.
Cartographic heat maps are criticised because they can be confusing. There are pros and cons to heat maps, just as there are to cluster maps, although my personal opinion is that they are an effective way of telling a story at a glance and so if the user does not know exactly what the colours and their gradients indicate, does this really matter so long as they correctly understand the theme and the overall picture being portrayed?
American sports multinational Nike has combined heat and cluster maps in my opinion to great effect in their Nike+ branding to advertise run locations within the UK. This is one of the best solutions to the problem of too many push-pins that I have come across. The points (or pins) are clusted and they uncluster as you zoom in. The heat map gives constant reference to the density of runs or runners in an area, or – seeing as I don’t know how it was made – I prefer to consider hot spots to relate to the ease of access or opportunity, rather than a pure density. I think the colours Nike have chosen are also commendable as they reflect the brand at the same time as working extremely well as an overlay to the Google base mapping.
Try it out for yourself here.
I was thinking about opening my blog with a post on heat maps and cluster maps a few months ago, so imagine my delight when I noticed how they are now being used globally in major events such as the African Cup of Nations. For many years heat maps have been a quick way of presenting information on a small, unobtrusive map at the edge of the screen. Throughout the Orange™ Africa Cup of Nations 2013 | South Africa, we were treated to some lovely bright heat maps everytime they displayed player statistics. They were used in conjunction with the value of the distance covered by a player during a match, to show the distribution of his running, i.e. his coverage across the pitch. As a cartographer, part-time gamer and a massive football fan, I found this addition to be valuable as well as it putting a smile across my face!
So there we have it. Cartographers and GIS users still argue about the value and design of heat maps – for which I would argue the most important cartographical consideration is the right base map (plug: OS VectorMap backdrop style 🙂 ), but I for one are pleased to see they have found their way into mainstream media and onto our TVs.