Filmfundr was the original genesis of Centiment, it aimed to fix film financing for people of color and filmmakers who struggle to access funds.
Centiment partnered with Scriptonomics, students from the NYU Center for Data Science, NYU Cinema Research Institute and Feilding University and IBM Research to assemble data science based, psychology based and sentiment based science built around understanding what content does and does not work, within film and all other current forms of content production – all driven by human thought based hypothesis.
This document discusses the Centiment.io and Scriptonomics content forecasting and performance prediction tools; the methodology of their development; and the validation of their results. These tools were originally designed to help screenwriters, first time filmmakers, film students, established directors, and individuals value and fund innovative and disruptive content — particularly content created by minority filmmakers who are female, of color, or queer. Centiment, in its current iteration, can be used in this capacity, but its scope has now been expanded to identify audiences for any form of video-driven digital, semantic and advertising content. We discuss the forces currently at play in the broader field of content forecasting, particularly those in the low budget content market. We also examine previous attempts at content forecasting, and the forces driving the need for the development of this capacity in technology.
First time filmmakers and content creators struggle to fund their content for a number of reasons: market demand, inexperience in the financing process, and complexity of distribution are just a few of many reasons. Fortunately, this problem has been addressed by digital distribution methods — but there is still no baseline comparison for understanding the market value of digital content produced by YouTube content creators, short film creators, and other independent groups. There have been extensive efforts to create such measurement systems for other content brackets; yet issues of socio-economic and societal significance complicate the process immensely.
The Centiment tool in its original iteration aimed to be the overarching solution for such users to value their content for content buyers and investors, and to function as a baseline for funding and understanding comparisons, thereby unlocking funds for this market.
The Scriptonomics tool provides deep content analysis of the text of a film script, extracting significant data points that allow content developers to assess its validity and improve writing accuracy according to key screenplay dimensions.
Using a cross-industry standard processes, the Centiment team explored data and metadata from 27 data sources relating to content, including IMDB, rotten tomatoes, box office mojo and a number of smaller blogs, social media sites and digital media data repositories. Scriptonomics applied their deep content analysis to hundreds of full length texts of film scripts from sources like IMSDB, DailyScripts, SimplyScripts and others, to construct a database of script features. Scritonomics used their analysis and comparison to historical data to provide creative feedback to screen writers and predict the financial performance of new screenplays. These screenplays were obtained from voluntary user uploads to the Scriptonomics website, offering a quantifiable methodology for versioning and greenlighting new content according to key script factors and dimensions.
Centiment White Paper Machine Learning and Emotional Content Prediction