Abstract
Entale is a small London-based startup bringing together product designers, engineers, data scientists, producers, and journalists whose goal is to transform the podcast listening experience through the Entale app. Entale is uniquely positioned to tackle podcast discovery, having developed robust pipelines for collecting, storing, and querying data from hundreds of thousands of podcast RSS feeds. In addition, Entale has released an in-house information retrieval system that extracts validated contextual content from podcast transcriptions in the form of linked named entities and external links added by the podcaster. Over time, Entale has also accrued a considerable volume of in-app user listening data.
The Entale challenge was composed of two interrelated goals. Firstly, we aimed to develop methods for capturing relationships between podcasts, incorporating information about the content discussed within them. Secondly, we worked to produce podcast recommendations, some of which utilised these inferred relationships – one key desired aspect of these recommendations was that they should allow ‘rabbit hole’ discovery experiences, where users could be pushed to further explore topics of interest.
Citation information
Data Study Group team. (2022, January 4). Data Study Group Final Report: Entale. Zenodo. https://doi.org/10.5281/zenodo.5818331
Additional information
PI: Jevgenji Gamper