A physical hive (left) streams three simulated sensor readings to its digital twin (right, dashed outline). The twin runs the same weighted health-score model every frame — nothing here is a canned animation, the score, status colour and recommendation all recompute live from the sliders.
tempScore = max(0, 100 − |T−34.5|·9)
humScore = max(0, 100 − |H−57.5|·2.6)
weightScore = 100, or 38 if W<16kg, or 68 if W>40kg
score = 0.42·tempScore + 0.33·humScore + 0.25·weightScore
score ≥ 78 → Healthy · ≥ 50 → Watch · else → Intervene
- Weight — total hive mass; a rising trend signals a nectar flow, a flat/falling one during active months signals a dearth or robbing.
- Temperature — brood-nest core sits near 34–35.5°C; deviations flag chilled brood, overheating, or a shrinking winter cluster.
- Humidity — 50–65% keeps brood and stored nectar healthy; too high risks mould, too low stresses larvae.
- Season — changes ambient bee traffic drawn around each hive entrance.
The single most urgent deviation (temperature, humidity or weight) is surfaced as one recommended intervention, exactly as a real precision-apiculture dashboard would.