Theoretical Foundation: Demand-Side Tracking
Historically, media ecosystems were analyzed by tracking what creators and networks published (the "Supply"). YT PoliSignal fundamentally rejects this approach, instead implementing Kevin Munger's Supply and Demand Framework (2024). We hypothesize that in modern algorithmic environments, creators do not lead audiences; audiences lead creators.
Therefore, our analytical engine stops optimizing for raw video counts and instead measures the intensity of audience Demand (engagement) to accurately predict what narratives will dictate the Supply tomorrow.
Engagement Heat Indexing
The core metric of demand is not "views" (which can be passive or algorithmically forced), but active friction. YT PoliSignal scrapes metadata across the ecosystem every 36 minutes, generating an Engagement Score where mathematically:
- Likes are weighted at a standard 1.0.
- Comments are weighted at 3.0, representing a significantly higher friction barrier of active audience participation.
By proactively comparing a video's active engagement density against its own channel's trailing 30-day baseline, we compute a Heat Index. A Heat Index of 5.0 indicates that an audience is reacting to a topic at five times the normal density, triggering a "Critical" alert on our metrics.
Audience Feedback Loops
To operationalize the theory that "Supply Follows Demand," the system actively hunts for Feedback Events.
When a specific broadcast hits the 75th percentile of historical engagement for a creator, the engine begins a 72-hour countdown. If the creator responds to this audience signal by rapidly publishing two or more follow-up videos exploiting the exact identical semantic keyword, a Feedback Event is natively logged.
This mathematically proves and objectively highlights when creators pivot their editorial coverage to capture audience rewards.
Affiliation Divergence Tracker
All 60+ monitored broadcasters within our dataset are rigorously stratified into two classifications:
Institutional Backing
Channels with explicit organizational ties to legacy media, advocacy groups, or major corporations (e.g., Fox News, Daily Wire, PragerU).
Personality-Driven
Autonomous content creators who exist entirely as their own brand and lack traditional external safety nets (e.g., Tim Pool, Glenn Beck).
By parsing semantic keywords and mapping them strictly to independent vs. affiliated networks, YT PoliSignal is able to output an absolute Divergence Index. This mathematically identifies narratives that are organically boiling up entirely from grassroots creators long before institutional outlets notice them.
AI Editorial Intelligence
Raw data is only as valuable as the synthesis derived from it. To provide a daily "bird's eye view," YT PoliSignal implements a local Large Language Model (LLM) pipeline.
Every evening at 20:00 local time, the engine aggregates the day's meta-signals—Pulse scores, divergence records, and breakout channels—and feeds them into a locally hosted instance of Ollama (Llama 3.2).
The AI acts as an objective editorial filter, synthesizing thousands of data points into a concise, high-signal briefing. This provides an automated layer of narrative detection that bridges the gap between quantitative tracking and qualitative media analysis.