Following an initial trial of Thoughtly’s technology, the pair have completed a detailed analysis looking at synopses and descriptions of programming alongside datamining to figure out how best to categorise individual programme titles.
Thoughtly’s flagship platform, Ellipse, is designed to map themes, generate summaries and identify anomalies in text. It was originally designed to assist researchers in academic institutions to draw insights from very large volumes of text, such as helping medical researchers identify unexpected anomalies in large clinical data sets or help scientific researchers navigate unstructured text for automated screening of ‘noisy’ data sets.
BBC Worldwide, the BBC’s commercial arm, is using the technology to identify themes that are under or over-represented in its content, to identify recurring and possibly unseen patterns across the various genres in it catalogue, identify which themes have grown and which have declined over the years and generally build a deeper understanding of its content with the objective of matching it with the most relevant audiences both for the BBC itself and for its client broadcasters, according to David Boyle, executive VP of insight.
“Through working with Thoughtly, and utilising machine learning, we are building a deeper understanding of our content so that it can be paired with the most relevant audiences for both the BBC Worldwide and our partners,” said Boyle.
“We are working together to answer some key analytical questions including: what are the recurring and elusive patterns that transcend the various genres in our catalogue? And how have themes evolved over the years – which have grown, declined, appeared and dissipated?”
Chase Perkins, founder and CEO of Thoughtly, said: “Thoughtly makes machine learning based research tools for non-technical professionals. If you want people to reap the utility of AI, you need to provide them with an immersive, app-like experience. We are thrilled to have partnered with BBC Worldwide’s Insight team as it continues to lead the industry and foster technology innovation.”
In-demand data has become increasingly important as a tool in commissioners’ and distributors’ arsenals, with the likes of Parrot Analytics providing services that suggest which shows are hot in certain territories and genres.