Launch readiness
Methodology
Destination recommendations blend cache-first pricing with seeded destination profiles, public-data overlays when available, and lens-specific weighting.
Pricing model
- Trip costs combine cache-first airfare with modeled hotel and daily-spend estimates. These are planning values rather than booking-grade quotes.
- Hotel and daily-spend estimates are heuristic and currently derived from PPP-style value signals, safety, trip length, and passenger count.
Current scoring backbone
- The current app starts from a destination reference table with PPP/value, safety, family-friendly, and retirement/livability proxies. Those fields are directional seed values in this build, not full live statistical feeds.
- Each lens then reweights that backbone using warmth, English-access, golf-scene, and a small number of lens-specific bonuses. Scores are for shortlist discovery, not definitive advice.
Current public overlays and next-source plan
- Today the app can load public golf-scene data from OpenStreetMap/Overpass and country-level English-access inputs from local files derived from public benchmarking sources.
- The source inventory in the repository maps the next credibility upgrades for each lens, including World Bank and WHO indicators, Open-Meteo climate normals, OSM amenity density, Wikidata, UNESCO, tourism-arrival data, and airport metadata.