Logo
ScienceStack
Table of Contents
Fetching ...
Sign In

WavePulse: Real-time Content Analytics of Radio Livestreams

Govind Mittal, Sarthak Gupta, Shruti Wagle, Chirag Chopra, Anthony J DeMattee, Nasir Memon, Mustaque Ahamad, Chinmay Hegde

TL;DR

WavePulse is presented, a framework that records, documents, and analyzes radio content in real-time, and demonstrates WavePulse's efficacy in capturing and analyzing content from radio livestreams sourced from the Web.

Abstract

Radio remains a pervasive medium for mass information dissemination, with AM/FM stations reaching more Americans than either smartphone-based social networking or live television. Increasingly, radio broadcasts are also streamed online and accessed over the Internet. We present WavePulse, a framework that records, documents, and analyzes radio content in real-time. While our framework is generally applicable, we showcase the efficacy of WavePulse in a collaborative project with a team of political scientists focusing on the 2024 Presidential Elections. We use WavePulse to monitor livestreams of 396 news radio stations over a period of three months, processing close to 500,000 hours of audio streams. These streams were converted into time-stamped, diarized transcripts and analyzed to track answer key political science questions at both the national and state levels. Our analysis revealed how local issues interacted with national trends, providing insights into information flow. Our results demonstrate WavePulse's efficacy in capturing and analyzing content from radio livestreams sourced from the Web. Code and dataset can be accessed at \url{https://wave-pulse.io}.

WavePulse: Real-time Content Analytics of Radio Livestreams

TL;DR

WavePulse is presented, a framework that records, documents, and analyzes radio content in real-time, and demonstrates WavePulse's efficacy in capturing and analyzing content from radio livestreams sourced from the Web.

Abstract

Radio remains a pervasive medium for mass information dissemination, with AM/FM stations reaching more Americans than either smartphone-based social networking or live television. Increasingly, radio broadcasts are also streamed online and accessed over the Internet. We present WavePulse, a framework that records, documents, and analyzes radio content in real-time. While our framework is generally applicable, we showcase the efficacy of WavePulse in a collaborative project with a team of political scientists focusing on the 2024 Presidential Elections. We use WavePulse to monitor livestreams of 396 news radio stations over a period of three months, processing close to 500,000 hours of audio streams. These streams were converted into time-stamped, diarized transcripts and analyzed to track answer key political science questions at both the national and state levels. Our analysis revealed how local issues interacted with national trends, providing insights into information flow. Our results demonstrate WavePulse's efficacy in capturing and analyzing content from radio livestreams sourced from the Web. Code and dataset can be accessed at \url{https://wave-pulse.io}.

Paper Structure

This paper contains 33 sections, 6 equations, 19 figures, 4 tables, 1 algorithm.

Table of Contents

  1. Introduction
  2. Framework and Data Collection
  3. Design of WavePulse
  4. A Dataset of Nationwide Radio Transcripts
  5. Analysis and Case Studies
  6. Case Study: Spread of a Political Narrative
  7. Case Study: Content Syndication Across Radio Stations
  8. Case Study: Presidential Candidates' Favorability Trends
  9. Related Work
  10. Discussion and Conclusion
  11. Supplementary Material for WavePulse
  12. Data Collection Pipeline
  13. Summarization
  14. Dense Embedding for Summaries
  15. Semantic Similarity Search with FAISS
  16. ...and 18 more sections

Figures (19)

  • Figure 1: Overview of WavePulse. It streams radio, transcribes, diarizes, classifies, timestamps and summarizes content on the radio, making available for analytics. We derive political trends, match claims
  • Figure 2: Coverage of Radio Stations. Each marker is an AM / FM station. We clubbed News/Talk/Business-News into "News/Talk", and Public-Radio/College/Religious/Others into "Other". Counts: "News / Talk" : 347, "Other": 49. For rest of the US plots, we will use above state labels as reference.
  • Figure 3: Samples of (Top left) JSON segments (Bottom Right) Corresponding Diarized Time-stamped Political News (Top right) Discussion, (Bottom Left) Advert., and (Bottom) Summary.
  • Figure 4: Occurrence of Neutral Reporting (51.0%), Debunking (36.3%) and Promoting broadcasts (10.4%) related to the 2020 Election narrative (2.3% were unknowns). We encoded number mentions in the size of bubbles and use identical scale throughout.
  • Figure 5: The "Syndication" Social Network among Radio. We do not show edges for clarity. Here, each marker is a station and color encodes its degree category.
  • ...and 14 more figures