Big Data & Analytics in the Digital Creative Industries
7 likes1,967 views
The document discusses the application of big data and analytics in the digital creative industries, particularly in film and television, highlighting various use cases such as predicting viewing habits, designing content based on user preferences, and improving marketing strategies. It also addresses the challenges of data collection, merging multiple sources, and interpreting data within the context of viewer interactions. The importance of leveraging big data for enhancing the viewing experience and the potential implications for creative industries are emphasized.
Big Data & Analytics in the Digital Creative Industries
- 1. Big Data & Analytics in the Digital Creative Industries Galit Shmueli 徐茉莉 Tsing Hua Distinguished Professor Analytics Humanity Responsibility Talk @ Dept. of Filmmaking , Taipei National University of the Arts, Oct 17, 2015
- 3. Get Big Data
- 4. Content and platform providers
- 5. Competitors, entrepreneurs, developers, researchers
- 7. Analytics: Use Big Data
- 8. 1. Predict viewing habits to provide user recommendations, display ads 2. Understand user preferences to design the next episode / series 3. Predict quality of experience to improve viewing experience 4. Identify target audience for new show to improve marketing strategies 5. Automate close captions creation & translation to improve user experience and expand market
- 9. 1. Predict viewing habits to provide user recommendations, display ads Google Play Music recommends playlists based on time of day, mood, and activities www.digitalmusicnews.com (10/2014) Search by mood, available time, plot preference… Semantic search (“movies with gangsters”) Input film you like, get rec’s for films, books, bands… Compares your ratings to others with similar “taste” Recommends similar films to movie you search for Evaluates your taste based on your movie ratings
- 10. https://gigaom.com/2012/09/09/can-breaking-bads-story-sync-get-viewers-to-give-up-their-dvrs 2. Understand user preferences to design the next episode / series
- 11. 3. Predict viewing experience to deliver the best results Improve quality of experience (QoE) Improve streaming quality
- 12. 4. Identify target audience for new show to improve marketing strategies www.fastcompany.com Universal hesitated about how to market their film. It was a "difficult title to compare to others"—was it more comedic? Did it resemble Rock of Ages more? Was it like Sparkle? Footloose? Or just a teen comedy? …Universal asked Analytics firm Fizziology to analyze how the film was being discussed on social media Before its release they discovered: Pitch Perfect was already morphing into a cult hit Universal changed their entire pre-release strategy. The film was released a week early, with a 1+1 ticket promotion Universal claims it was the most extensive screening campaign in the studio's history
- 13. 4. Identify target audience for new show to improve marketing strategies Netflix used big data for color-design of the cover
- 14. 5. Automate close captions creation & translation to improve quality of experience (QoE) and expand to new markets Speech recognition Natural language processing
- 15. Competitors, entrepreneurs, developers, researchers Analytics: Use Big Data
- 16. The social interactions between viewers while watching TV https://youtu.be/ipfsuKwQYu0
- 17. Let’s see Two Cases of Big Data and Analytics in Film & TV
- 20. Even if you have the data, it isn’t easy to get recommendations right
- 21. Satyamev Jayate Popular Social TV Show in India
- 26. Study (based on Season 1) by Galit Shmueli & Sandeep Khurana Indian School of Business, 2013
- 27. Publicly Available Data Television Viewership Ratings Simulcast Regional languages Radio talk shows Blogs Print Media Social Media YouTube Facebook Twitter Website Search traffic trends (Google)
- 28. YouTube
- 29. Google Trends Daily volume for top regions/cities • Download CSV • Missing for low volume • How is a region defined? • Other breakdowns?
- 30. Wikipedia
- 31. Looking at multi-source data
- 32. Challenges for Outsiders Technology • Data collection • Merging multiple sources Data availability • Archiving policy (Twitter) • Provider-determined (Google Trends) • User-determined (YouTube) Meaning of data • What is a YouTube “view”? Google Trends “State”? • Missing data? Merging data sources • Level of aggregation • What is a record? Long term collection: Changes in data dissemination policy
- 33. Analytics for your own YouTube content are richer
- 36. There’s plenty of information also on other users’ content Data YouTube gives us directly Data we can scrape from YouTube pages www.youtube.com/watch?v=YphFNAVAdpE
- 39. DNA of a Song gender of lead vocalist, prevalent use of groove, level of distortion on the electric guitar
- 41. Google Developers On the challenge of YouTube video recommendations
- 43. Use humans to label a clip/movie/music: • Tag “best moment” • Label their emotions & reactions (sad, laughing, scared, surprised…)
- 44. Alternatives to Amazon Mechanical Turk (Chinese language websites?)
- 45. Analytics: Finding Patterns in the Data (using machine learning) Recommendations and “what goes with what” Segmenting videos/users/etc. Predicting user behavior After tagging a large set of songs, videos, users… You can use machines for:
- 46. Competitors, entrepreneurs developers, researchers Can Get Big Data Use Big Data Analytics
- 48. “I'm a data miner, but I hope creative industry doesn't get corrupted by data mining… I have read an article recently in Wired about a music hit made basically based on data.” (post on an Internet forum) Andrew Leonard at Salon.com, considers some of the implications of TV in the era of Big Data A creator is always looking ahead, never looking back
Editor's Notes
- #9: User preference: Breaking Bad polls audience about plot in previous episode Quality of experience: Netflix collects end-user data on physical location and buffering delays and bitrate. Range of devices, with different connectivity. Target audience: Netflix uses for cover design for House of Cards; Universal uses for release of musical film Pitch Perfect
- #10: http://www.cnet.com/news/top-10-movie-recommendation-engines/
- #11: User preference: Breaking Bad polls audience about plot in previous episode http://www.fastcodesign.com/1671893/the-secret-sauce-behind-netflixs-hit-house-of-cards-big-data
- #12: Quality of experience: Netflix collects end-user data on physical location and buffering delays and bitrate. Range of devices, with different connectivity.
- #13: Target audience: Netflix uses for cover design for House of Cards; Universal uses for release of musical film Pitch Perfect http://www.fastcompany.com/3024655/pitch-perfect-and-how-analytics-are-transforming-movie-marketing
- #14: Target audience: Netflix uses for cover design for House of Cards; Universal uses for release of musical film Pitch Perfect http://www.fastcompany.com/3024655/pitch-perfect-and-how-analytics-are-transforming-movie-marketing
- #15: Target audience: Netflix uses for cover design for House of Cards; Universal uses for release of musical film Pitch Perfect http://www.fastcompany.com/3024655/pitch-perfect-and-how-analytics-are-transforming-movie-marketing
- #19: http://www.forbes.com/sites/robcain/2015/06/16/announcement-of-chinas-netflix-may-be-the-death-blow-for-netflix-in-china/
- #48: Salon.com is a US news website, that focuses on US politics from a liberal perspective, and on reviews and articles about music, books and films
- #49: http://www.salon.com/2013/02/01/how_netflix_is_turning_viewers_into_puppets/