PJQv1.2

Seismograph

Seismograph

The Seismograph is the primary analysis tool on the PJQ platform. It allows you to scan the audience's reaction to a specific video and receive a structured verdict.

Running an Analysis

1. Navigate to the Seismograph tab.

2. Paste the URL of the target video into the input field.

3. Click the run button.

Processing time depends on the volume of comments and the current load on the computing nodes. Usually, generating a verdict takes anywhere from a few seconds to a few minutes.

Measurement Modes

Before running, you choose a measurement mode that trades speed and cost for stability. A verdict reads a *sample* of the comments, and a larger sample gives a more stable result — so the mode is your main lever over confidence.

  • Scout — a quick, rough read on a small sample, without QA escalation. One credit. Best for noisy content (entertainment, shorts) or a fast first look, where pinpoint accuracy isn't the goal.
  • Standard *(recommended)* — a balanced random sample with QA escalation enabled. One credit. The battle-tested default; the right choice for almost everything.
  • Deep — a large sample with QA, for the most stable reading. Two credits. Use it for divisive or high-stakes topics where you want the Mood-Net to hold steady on re-runs.
  • QA escalation re-checks low-confidence labels against a second model, trimming misclassification. It is on in Standard and Deep, off in Scout. Each mode shows its sample size and credit cost up front, so there are no surprises on your balance.

    What You See on the Dashboard

    Once the analysis is complete, the Seismograph will display the verdict dashboard. The dashboard consists of three logical blocks:

    1. Global Indexes: Three gauges (Mood, Controversy, Parasitic) that provide an instant read on the "temperature" of the video.

    2. Stance Distribution: A chart showing the exact percentage of comments across each of the 7 categories (ranging from meaningful support to off-topic).

    3. Example Quotes: 2-3 real comment examples for each detected category. This allows you to quickly verify the accuracy of the assessment without having to read the entire thread.

    Scanning Features

    The Seismograph analyzes a representative sample of comments. The tool automatically normalizes the data to prevent bot attacks (e.g., mass posting of identical messages) from skewing the final percentages of the meaningful categories.