METHODOLOGY

DSTI Methodology

Last updated: 2026-09-07

Positioning: Trend Score, not Index

DSTI (DigitalScope Trend Score) is a relative trend score. We deliberately avoid the word “index” early on: a score is explainable and iterable, avoiding premature absoluteness. Weights, formula versions and baseline calibration are published on this page.

Scoring factors (t1 · heuristic-v1, the current implementation)

The current scoring engine is a deterministic five-factor weighted sum (version heuristic-v1): baseline 20 + source tier (T1=18/T2=10/T3=4) + multi-source confirmation (2 sources +14 / ≥3 sources +22) + recency decay (<6h=20 / <24h=14 / <72h=8 / older=0) + AI impact-analysis confidence (0–30; zero — never inflated — when enrichment is pending). Total clamped to 0–100. Without AI, confidence is fixed at 0.35 (heuristic tier). The planned Volume/Momentum/Engagement/Breadth system is a future version and is not claimed here until implemented.

Baseline and trend semantics

Trend momentum = the relative change rate of per-category signal volume in the current 7-day window versus the previous 7-day window; when the previous window is zero and signals appear, momentum takes the full value 1 (semantics: topic emerging from nothing, not an explosive ratio). Calibration uses observation data before launch; until then the site honestly shows empty states — never a fabricated score.

Limitations

DSTI measures attention and activity, not technical merit; it reflects our curated source allowlist and carries coverage bias.

DSTI is for trend reference only and does not constitute investment advice.