The BB-8 droid from "Star Wars: The Force Awakens." Credit: Disney | Lucasfilm Do the robots of "Star Wars" or the genetically-engineered dinosaurs of "Jurassic Park" have any influence on real-world scientific research? Many scientists and engineers can probably list their favorite science fiction films that inspired their careers and work. But data science could help dig deeper into science fiction's influence within the thousands of research papers found in online databases. This idea of identifying influential sci-fi films in research grew out of a conversation I had with Sparrho, a UK startup founded by ex-scientists who wanted a better tool for finding relevant research. Since 2013, Sparrho's founders have developed a search engine for scientific research that uses machine learning algorithms to tailor search result recommendations to each user's individual interests. Sparrho also has access to the online database of the British Library, the UK's national library, through a special partnership. That means the online service can access research paper abstracts going back to the 1890s.
Finding the Top Contenders
The Sparrho team graciously agreed to trawl through the available data in search of possible sci-fi influences. Katja Bego, a data scientist at Sparrho, started out with a list of 40 popular science fiction films selected through a combination of moviegoer ratings, movie critic ratings and box office gross. Films that appeared in the top 100 lists in at least two of those categories made the cut. Bego then ran searches for the movie titles among the abstracts and titles of research papers in the British Library database. She also manually checked the results to only count papers that explicitly referenced and discussed the films. Articles that merely mentioned a film title in passing did not count (e.g. "Other blockbusters, like movies A,B & C, were less profitable.") Some films such as "The Matrix," "Terminator" and "Alien" generated too many search results to filter manually. In those cases, Bego ran an approximate search by looking for common keywords related to the films. Here is the list of popular sci-fi films that turned up at least two relevant search results. Not all the relevant search results represent physical science research; they also include film studies or social science results. (The film list also includes the dystopian film "Soylent Green" that isn't necessarily among the most popular films, but was thrown in as a wild card.)
Star Wars (18 relevant papers)
Jurassic Park (11 relevant papers)
2001: A Space Odyssey (9 relevant papers)
Blade Runner (8 relevant papers)
Minority Report (4 relevant papers)
Back to the Future (3 relevant papers)
The Matrix (2 relevant papers)
Soylent Green (2 relevant papers)
This list overlaps to some degree with informal polls of researchers done in the past. In 2015, a Popular Mechanics poll that included "dozens" of scientists generated a list that included "2001: A Space Odyssey," "The Matrix," "Jurassic Park" "Blade Runner" and "Star Wars." A 2004 poll of scientists by
