It is sort of touched on later in the article, but the first graphic, showing the number of native speakers for each language, isn't very useful. The statistic that better measures the utility of knowing a language is the total number of people with working knowledge of the language, which would boosts the numbers of English, French and Russian dramatically.
I wonder if a better statistic would be number of speakers with working knowledge of a language, weighted by GDP per capita of the countries in which those speakers live. Of course, that assumes the purpose of learning a language is to do business, which is not necessarily true.