Is it a good day to run?

The score is a 0–100 composite built from six numbers that actually matter to a runner: feels-like temperature, rain, PM2.5 air quality, wind, UV index, and a dew-point humidity deduction. The weighted subscores are: temperature 30%, rain 25%, air 23%, wind 13%, UV 9%. Humidity then comes straight off the blended score, up to 35 points, scored on dew point because relative humidity misleads on its own: it starts to bite near dew point 16°C, reads muggy at 18°C, oppressive at 21°C and brutal at 24°C. UV uses the WHO scale: ≤2 low, ≤5 moderate, ≤7 high, ≤10 very high, 11+ extreme. UV doesn’t cap the score (sunscreen + a cap mitigate it), but a midday UV 10 in July still drops it by ~4 points and pushes sun-protection into the wear list.

The score is graded against the season. A 23°C dawn in late July is close to the best July offers, so it scores well for July, and the verdict says so while still calling out the heat and humidity in absolute terms. The same dawn in October would score poorly. Safety ceilings (heat, air, dust, wind, rain) never shift with the season.

You can tune any of these to your own tolerance via the Personalize the scores panel near the top of the page: heat / cold bail point, PM2.5 bail point, UV bail point, humidity, rain. The defaults match the weights above; your prefs are stored on this device only.

Five things will cap the score regardless of how good the other inputs read, because they reflect a physiological reality the runner can’t muscle through:

• Heat   Shade feels-like above 24°C caps the score at 80, above 27°C at 63, above 30°C at 49, above 33°C at 36, and above 36°C forces a Skip.
• Air quality   PM2.5 above 35 μg/m³ caps the score at 78. Above 50 caps at 60. Above 75 forces a Skip.
• Wind   Sustained wind above 5.5 m/s caps at 78. Above 7.8 m/s caps at 55. Above 10.5 m/s forces a Skip, because Han River paths and bridge decks get unsafe.
• Asian dust (황사)   Coarse-particle event in progress caps at 45 even when PM2.5 alone reads moderate. KF94 mask required.
• Humidity   Dew point at or above 24°C caps the score at 55, at or above 26°C at 45. Evaporative cooling stops working, whatever the air temperature.

For warm hours the score runs on the shade feels-like, a humidity-driven heat index and the same convention KMA and Naver use, so the heat ceiling, the reasons and the number on the card all quote one figure. Cold hours keep the wind-chill feels-like instead, since that is what actually bites.

Weather forecast data comes from Open-Meteo (ECMWF + DWD ensemble). Current air-quality readings come from AQICN’s aggregation of real AirKorea (한국환경공단) station readings, so the PM2.5 number for each zone is from an actual Korean ground station near that zone, not a model grid cell. The next couple of days use the same air data as the hour-by-hour view: Open-Meteo’s air-quality forecast (AQICN doesn’t publish hourly numbers), anchored to the live ground reading so the near-term hours track what’s actually in the air. Days further out, past that anchored window, fall back to AQICN’s daily PM2.5 outlook.

The page caches each upstream fetch for ten minutes at the Cloudflare edge, so reloads aren’t hammering the APIs.

Score ladder: 88+ as good as Seoul gets. 78–87 is a strong run window. 65–77 is solidly workable. 50–64 is manageable but compromised, keep it easy. Below 50, indoors is usually the smart call.

The scoring rules are tuned for an average-fit adult running 5–15 km in Seoul. They aren’t medical advice. If you have a respiratory condition, are pregnant, or run with kids, default stricter on the air-quality side.