Temporal Cascade Influence Score (TCIS)

What it measures

TCIS is a timing signal. It asks:

• Did stronger communities speak first?
• Is momentum growing more in stronger communities or weaker ones?

So CCSC tells you "are communities agreeing?", while TCIS tells you "how the story is spreading over time."

Inputs

TCIS uses:

• The time each community first mentioned the ticker.
• Mention volume by community.
• Community trust scores.

How it is computed

TCIS has two parts:

1. Origin quality:

   • Look at the first few communities to mention the ticker.
   • Give earlier mentions more weight than later ones.
   • Communities with higher trust scores contribute more.

2. Momentum quality:

   • Compare mention growth in higher-trust communities vs lower-trust communities.
   • If momentum is stronger in high-trust communities, this part is positive.

Then both parts are blended 50/50 into one score.

Output + interpretation

• Higher TCIS: the signal is being led by stronger communities and/or momentum is concentrated there.
• Lower TCIS: the narrative is mostly led by lower-trust communities, or leadership is unclear.
• Near zero: mixed timing signal.

Edge cases/default behavior

TCIS stays stable under sparse data:

• No signals: returns neutral (0).
• Missing first-mention timing: origin part weakens, but momentum part still works.
• The momentum formula includes a guard to avoid divide-by-zero.

How Rebuzz uses it

TCIS is recalculated daily and fed into the composite Rebuzz score.

It adds a "who led, and how fast" dimension that pure sentiment averages cannot capture.