Earlier this year the DOAJ cleaned its index and removed about 3,300 journals, this explains why the numbers for 2016 differ so much from the heatmaps of 2015 and 2014.
Please note that the map has only illustrative value and of course it is obvious that it needs to be contextualized with other information to allow profound conclusions. Anyone who wants to have more detailled or granualar information is invited to take the data available on the web and to build heatmaps that visualize for instance the number of Open Access articles published per country. Of course it may also be useful to take other indicators into account as consumer price index, expenditures for research or the number of scientists.
Although I am very well aware of all these limitations of the heatmap published here unfortunatley I do not have the time to collect the data and build these maps – it is up to the rest of the Open Access community to do so if it thinks better maps are needed.