Just a short remark: the most statistically explicative parameter for Wikipedia activity is not the number of internet connections but GDP (except for English and Chinese projects which exhibit singular behavior). Perhaps you could retry the analysis using GDP and find some more countries where chapters are achievable. Sorry for not providing references. This comes from a not yet published research work that applies reasonable hypothesis to transfer more than 20 parameters from country data into language data and then apply statistical methods to search for correlations between those data and size of Wikipedia projects.