Skip to main content

S6. External Signal Fusion (Mitsukoshi)

Published 2026-05-14Updated 2026-06-301 min read

1. URL · Personas​

  • /signals · P2, P3

2. 4 External Signals (Mitsukoshi-specific)​

TypeSource
SocialXiaohongshu · Dcard · Instagram · Japanese Twitter · Hong Kong Facebook
WeatherCentral Weather Administration of Taiwan (中央氣象署)
FX (multi-country)JPY/USD/HKD/SGD ↔ TWD daily
TourismTaiwan Tourism Bureau foreign arrivals (by nationality)

3. Processing Pipeline​

  1. Aggregate first-party GMV by day (store, BU)
  2. Join with 4 external types
  3. Correlation and lag analysis
  4. Visualization: time series + heatmap + tourism annotation

4. Output UI​

  • Dual-axis time series (external + first-party)
  • Category × external correlation heatmap
  • Lag analysis (lag 0–14 days)
  • FX fluctuation annotation

5. Demo Scenarios​

  1. Weak yen -10% → Japanese duty-free -25% (lag 7 days)
  2. "Lifestyle" keyword vs first-party SKU
  3. Taiwan foreign arrivals +20% → Mitsukoshi foreigner revenue +15%