Top 10 Programming Languages

Building a Better Ranking

When IEEE Spectrum wanted to update its 2011 list of the top 10 programming languages, Nick Diakopoulos was a natural choice for the project. Diakopoulos is a leader among the new breed of computational journalists. He couples a reporter’s story sense with programming expertise in big data, visualization, and interactive technologies, and he is a fellow at the Columbia University Journalism School, in New York City.

Diakopoulos says the initial hurdle was just figuring out which programming- related data sources had useful information that could be extracted via an application programming interface, or API. “Spectrum’s editors and I came up with 10 to 15 sources; then I had to go through and figure out which ones were realistic,” he says. The sources included Topsy’s database of tweets, open-source repository GitHub, and the IEEE’s own Xplore archive of technical papers.

Once the sources had been identified, Diakopoulos began working out how to combine the data from them and formulate rankings. “We let the notion of transparency guide the design, so that users could play with the weighting of the data sources to come up with their own rankings,” he says. After about three iterations of the design, the result is the general-purpose ranking you see in Dataflow (“The Top 10 Programming Languages,”). For the interactive application, go to the online version of the article. It lets you customize the ranking for your own particular interest, whether it’s finding out what language skills are hot in the job market or which languages are dominant in the mobile sector.