Brian and I recently published a paper on the arxiv:, earlier this month it was accepted into the ChASM workshop at WebSci. Here’s the abstract:

A number of attempts have been made at estimating the amount of box office revenue a film will generate during its opening weekend. One such attempt makes extensive use of the number of views a film’s Wikipedia page has attracted as a predictor of box office success in the United States. In this paper we develop a similar method of approximating box office success. We test our method using 325 films from the United States and then apply it to films from four foreign markets: Japan (95 films), Australia (118 films), Germany (105 films), and the United Kingdom (141 films). We find the technique to have inconsistent performance in these nations. While it makes relatively accurate predictions for the United States and Australia, its predictions in the remaining markets are not accurate enough to be useful.