Worldmapper has become a popular repository for a slightly different visualisation of some of the most relevant issues of our time. We get a large number of emails, often asking very similar questions, so I want to take the opportunity to take you behind the scenes of worldmapper and answer some of the most frequently asked questions about the project. Let me start off this time with some background information about us and about how up-to-date the worldmapper maps are. 

The worldmapper website was launched in 2006 with a first series of about 300 maps being included initially. These have now grown to over 700 different maps, with the series about death, religion and language being the most recent additions that we added. Currently there are no major revisions in work as the capacities that we can put into worldmapper are much more limited than in the first year of the project, where there were many people doing really great work.

Instead we started looking into new possibilities to move the worldmapper visualisations further than showing nation-level data only. The release of the world population atlas as an integral part of worldmapper is the first result of these efforts. But we also continue creating the odd worldmapper map looking at current events and issues, such as our look at the recent Malaria-related death figures in a map that we created for UNICEF.

 

Many of the topics that are covered by worldmapper rely on these efforts, and many of the data that we used to create the worldmapper maps basically build on statistics from the census rounds almost a decade ago. Newly released data often comprises of intercensal estimates - meaning that the census information is projected to the expected changes in a given time from what we observed before. And for many countries such more regular estimates don't exist at all, making it hard to compile consistent global data sets that reflect an honest representation of the existing conditions.

The worldmapper maps need consistent and complete datasets, because the underlying mapping technique does not allow for any gaps in the data that we use without resulting in a serious misrepresentation. On a normal map we can map no data by showing a country in grey - on a worldmapper map no data would result in a country to disappear and thus give a very false impression. That is why we have to put a lot of effort in closing the gaps wherever they occur.

If you are one of those who have a (sensible) suspicion of everything that says statistical evidence, then you may be aware of the obstacles that such an estimation of data requires. To be as transparent as possible, all original data sources and all adjustments that we applied to the data before creating a map are documented in the supplementing material that we provide for every map on the worldmapper website. This is also a great source for further work with the data and for using it as a resource for teaching or for your own statistical analyses. 

One last thought about the maps themselves: Technically worldmapper maps are also known as cartograms. There are many ways to draw cartograms - the older ones amongst us may remember for instance the rectangular depictions of the world's population that were a common feature in many school atlases. The worldmapper cartograms appear more map like because they are based on a very recent cartogram technique developed by two physicists (Michael Gastner & Mark Newman) who used a concept of elementary physics create the cartogram transformation. This allows the individual country shapes to appear more like they look on a normal map (always depending on the degree of distortion of course).

What the worldmapper (and other) cartograms show are nothing else than cartographic versions of a pie chart. The map shows the share of each country on the topic that is shown in the particular map, but in a very elegant and certainly much more understandable way that those slices of a pie chart can ever do for such a large range of numbers. A ‘normal’ map thus is the same as a land area cartogram, as all countries are sized according to their land area shares. Which is the more valuable visualisation – a map, or a pie chart?

If you keep the idea of a pie chart in mind, then this also explains why often don't even need to create constant updates of a map. Even if we had very recent statistics, the shares between the countries change very slowly. Take GDP as one measure for the wealth of a country. With all the talks about the economic crisis you may expect the GPD cartograms to look very different from only a few years ago - but if you think again, you will realise that these are global trends, and than a slowdown in economic activity affects all countries. Thus the shares between the countries don't change that significantly in a shorter period of time, as growth or decline is a global phenomenon (similarly you can see the same phenomenon for the economy of the European Union, that we recently look into at the worldmapper blog. Only when we start to look at longer time periods, such as the differences that happen within a decade, we start to see these gradual changes slowly becoming visible in the cartograms. 

The two GDP maps of 2000 and 2010 show, how slow the global changes in the distribution of wealth affect the map appearance, despite the total changes that happenedin this decade. So be confident when looking at many of the worldmapper maps, even if some of them are two, three, four or even five years old. Most of the maps still provide very good representations of the correct shares between the countries and help to understand how the different parts of the world relate to each other.

Enough said about worldmapper this time. Perhaps there is some more time and space later this month to shed more light on some other elements of the project.

Ben Hennig