Using Data, Networks and Complexity to Study Trade, Aid, Economics

Following the announcement of a new way of looking at how interest in development projects might have moved around Africa over time (see this blog post) I’ve produced a more subtle version with one particularly interesting twist:

One of the problems with comparing countries against each other is that some are huge and some are tiny. (Note that there are hundreds of problems with comparing countries with each other, and this is merely the easiest to get around!) By adding population data from the UN’s Department of Economic and Social Affairs we can try to ‘account for’ the fact that larger countries will inevitably generate more Googling interest.

How does this work? Well, we simply assume (for now at least) that the size of your population has a linear effect on your Google page count: this means that if country A has a population twice that of country B then, all things being equal, we expect country A to have twice as many Google pages about it. The page count numbers are then adjusted so that all the countries have the “same population”. So small countries have their page counts increased and large countries have their page counts decreased, according to just how big or small each country is. This is clearly a gross simplification, but it gives a better idea of the relative Google interest over time than not adjusting for population at all.

Google result count by project type and year

click here for the interactive version.

Post a comment