Pivotal at NAB 2015 Solutions Workshop
The Solutions Workshop at the EMC Isilon booth at NAB 2015 was sponsored by Tamed Technology – and featured workshops, presentations and Q&A sessions from a number of partners and vendors – including Pivotal. Pivotal is part of the federation of strategically aligned businesses associated with EMC.
Noelle Sio – Principal Data Scientist at Pivotal – presented a very interesting case study examining the power of data science applied to help identify the programme production decisions that have the most effect on ratings.
The Pivotal Data Labs team set to work to try and identify what makes viewers tune-in and tune-out of programmes they are watching. Whilst a large amout of data is available, much of it is unstructured – and even presented in inconsistent forms. A large part of the work therefore was to normalise and clean-up the data.
Noelle talks here about her and Pivotal’s approach to the problem.
Noelle Sio. Principle Data Scientist, Pivotal.
Using the tools and techniques at their disposal, the Pivotal team were able to build an open solution – thereby setting the ground-work for future analaytics – which would scale both for the number of programmes and the number of broadcasts.
Many will instinctively say that the format of the show, and the strategy for commercial breaks has the highest impact on viewership. What the team found was that a number of other variables are important too, including:
- Speaker characteristics
- Number of people on the screen at a time
- Broadcast topics
Those working with social media networks are familiar with data analysis – and using data to drive decisions about content. However, this is not common in broadcast, and the work by Pivotal is an interesting extension of the science.
Much more information can be found about the case study at Pivotal’s website.