How Data Science Is Used Within the Film Industry
As Data Science is becoming pervasive across so many industries, Hollywood is certainly not being left behind. Learn about how Big Data, analytics, and AI are now core drivers of the movies we watch and how we watch them.
By Frankie Wallace.
There are countless factors at play in filmmaking, from determining production costs to developing targeted marketing campaigns. Data science is involved in practically every step of the process, and professionals who work in data science can learn many things from the film industry.
Streaming services are at the forefront of the data science revolution. Production companies, including Amazon, Hulu, and Netflix, analyze patterns in big data to determine the types of content they create and make personalized viewing recommendations. In this way, data science can aid the art of producing and marketing entertainment at levels never before seen.
The field of data science also pops up as meaty subject matter in a variety of films. The stories of real-life innovators such as Alan Turing and John Nash have been turned into major films in recent years, living alongside fictionalized tales that use predictive analysis, machine learning, and AI as central plot themes.
Society’s fascination with the implications of data science indicates that more films on the subject are sure to come. Further, production companies will continue to use the technology to better understand individual viewing habits and preferences to create content that appeals to the masses.
Film Success Metrics and Relevant Data
Image Source: Pexels
Technology can inform filmmakers how they should produce and market any given movie. From casting decisions to even the colors used in marketing, every facet of a movie can affect sales. Using technology, we can predict customer preferences and determine how to optimize content to reach its maximum potential.
Predicting what audiences want from a film almost guarantees that film’s success. In 2018, 20th Century Fox, which was acquired by the Walt Disney Company this year, released a paper outlining how it analyzes the content of movie trailers using machine learning. Data collected in the process is used to compare trailers and predict what other films might interest those who watched a particular trailer.
20th Century Fox used Google servers and the open-source AI framework TensorFlow to create Merlin, an “experimental movie attendance prediction and recommendation system.” In Merlin’s trial run, the tool analyzed the trailer of “Logan,” an origin story of the superhero Wolverine, to predict other movies that “Logan” viewers might be interested in. Of the 20 predicted, 11 were correct.
The top five actual movies were all in the predicted list: X-Men: Apocalypse; John Wick: Chapter 2; Doctor Strange; Batman v. Superman: Dawn of Justice; and Suicide Squad. Generally, the audience was looking for a superhero movie that featured a “rugged male action lead.”
While its data interpretation wasn’t perfect, Merlin is a prime example of the evolution of software development over the last decade. For programmers to better concentrate on improving AI algorithms, future software development must include time-saving measures designed to reduce time spent on menial tasks. As AI is designed to focus on a single task, it’s an ideal starting point in improving the accuracy of data analysis within programs.
The Role of Big Data in Analytics
When big data first hit the scene around 2010, it effectively changed the methods used to turn data analytics into useful insight and profit. Big data is often externally sourced, using information drawn from the internet, public data sources, and more to make more accurate predictions. In the entertainment industry, big data can be used to provide a personalized user experience and reduce churn rates among streaming site audiences.
With a seemingly endless array of movies and television shows for users to choose from, retaining viewers is of paramount importance to streaming services and film production companies. A high churn rate indicates that a company is doing something wrong, and when combined with machine learning, big data can help companies identify problem areas.
Among streaming services, the user interface plays an important role in viewer retention. If viewer recommendations are inaccurate, for example, it could lead that viewer to turn to other platforms for entertainment. Streaming services are well aware of the importance of a positive user experience.
To keep viewers engaged, Netflix developed and continues to improve upon its adaptive streaming algorithms to optimize streaming quality and create a personalized user experience. The streaming giant adjusts the audio and visual quality of the media to optimize the experience. They also use predictive caching to allow a video to play faster or at a higher quality. For example, if a viewer is watching a series, the next episode will be partially cached.
Meanwhile, the recommendations are based on both explicit and implicit information. “Explicit data is what you literally tell us: you give a thumbs up to The Crown, we get it,” Todd Yellin, Netflix’s vice president of product innovation, told Wired. "Implicit data is really behavioural data. You didn’t explicitly tell us 'I liked Unbreakable Kimmy Schmidt', you just binged on it and watched it in two nights, so we understand that behaviourally. The majority of useful data is implicit."
And, if its profits are any indication, Netflix algorithms are a resounding success: Since 2015, Netflix profits have grown more than 30%, with revenue amounting to $16.614 billion annually.
Predictive Analytics in the Film Industry
Image Source: Pexels
The implications of Merlin and similar programs for predictive analytics are wide-reaching, but a larger subset of data is needed to find accurate patterns. Over the last several decades, researchers have collected data on thousands of movies and television shows in search of viable predictive indicators. Correlations have been found in numerous categories, including character types, plot complexity, star power, budget, and “buzz,” or the social chatter and marketing presence surrounding a particular film.
Buzz is notable in the sense that information on the phenomenon can be gained from numerous sources, such as social media and critical reviews. The buzz surrounding a film is only a small piece of the larger analytical picture, however. Data analytics must be used at every life cycle stage of the movie, from development to post-production and distribution.
Predictive analytics can help producers, production companies, and executives to inform strategic decision-making, predict trends, and better understand viewer habits. Informed decision-making is imperative to the film production process, and acquiring high-quality, highly usable data is key to customer retention and profits. Data scientists should take note of the myriad ways that the film industry utilizes predictive analysis and big data and bring that knowledge to other industries and business settings.
Bio: Frankie Wallace is a freelance writer from the Northwest who contributes to wide variety of blogs online. Wallace currently resides in Boise, Idaho and is a recent graduate from the University of Montana.
- Which Movie Sequels Are Really Better? A Data Science Answer
- Building a Recommender System
- The Infinity Stones of Data Science