Deciding which language is “most useful” is a question of what you are trying to do. The most logical answer is that the most useful languages are those that are spoken, and connected to, as many other languages as possible. To figure out the answer to this question, Shahar Ronen decided to crunch all available data from books (based on 2.2 million translations of books into 1000 languages), twitter (based on 550 million tweets by 17 million users in 73 languages), and Wikipedia (tracked edits in up to five languages done by editors, carefully excluding bots) to see what was being translated into which languages.
Shahar Ronen, a Microsoft program manager whose Massachusetts Institute of Technology (MIT), is a bilingual Hebrew-English speaker from Israel. The project began when he told his MIT adviser, César Hidalgo (who is himself a Spanish-English speaker), about a book written in Hebrew which had yet to be translated into English. He summed this thought up by saying; “I was able to bridge a certain culture gap because I was multilingual,” and began thinking about how to create worldwide maps of how multilingual people transmit information and ideas.
Along with co-authors from Harvard and abroad, Ronen created a visual model to display the millions of pieces of translational data. Of course, more detailed information can be obtained from the study that resulted from this data, in which English appears to play the most global role as an intermediate (or “bridge”) language. You can look at and play around with the visual representations of the data here, but I will attach the “primary” images here:
Here is the data based on the 550 million tweets by 17 million users:
And lastly, here is the data based on the Wikipedia edits:
All of the graphics use the same legend, so this should make interpreting the graphics easier: