How the Siesta Score is calculated
The Siesta Score measures night noise from objective data. It uses no reviews and no hotel self-description, except for the sleep facts, which are labelled as declared.
Where the noise data comes from
From the strategic noise maps city councils are required to publish under EU law (Directive 2002/49/EC), Lnight indicator. We take the highest value within 20 metres of the hotel — its most exposed façade. These maps model road traffic.
From 0 to 10, per the WHO
The score does not come from a scale we invented. It comes from the dose-response curve the World Health Organization published in 2018 for sleep disturbance from road traffic, which relates Lnight to the share of people whose sleep that noise highly disturbs. We anchor 10 at 40 dB, the WHO night guideline, and 0 at 70 dB. So a 5 is not «average»: it means one in nine people would have their night ruined. Each hotel page shows that percentage.
The formula
Total Lnight → score: WHO curve · 40 dB → 10 · 70 dB → 0 % of people whose sleep is highly disturbed = 20,8 − 1,05·L + 0,01486·L² Total Lnight adds the sources in energy (10·log10): · road traffic, from the official map · aircraft, where the map publishes them · nightlife: 40 + 10·log10(venues within 250 m), −3 dB on weeknights − 1.5 major road < 50 m − 0.75 < 100 m − 1.0 surface railway < 100 m − 0.5 < 200 m − 0.5 hospital < 200 m − 2.0 / − 1.0 runway of ≥1,200 m within 1.5 / 3.5 km, where the map does not measure aircraft cap on the distance corrections: − 3.0 + 1.0 declared soundproofing + 1.5 verified + 0.5 double glazing (max +1.5)
How precise each city is
Cities do not publish at the same level of detail. Madrid publishes continuous values on a 5-metre grid, Barcelona on a 2-metre grid. Berlin measures point by point at the façade, the best data there is. Paris, Valencia, Bilbao and Malaga publish 5 dB bands, so scores there carry a ±2.5 dB margin and move in steps. Age varies too: almost all come from the 2022 European round, but Malaga's map is from 2013, the oldest we use. All cities use road traffic noise only, even where rail and industry are published too: otherwise cities would not be comparable and railways would count twice, since they are already penalised by distance.
Nightlife areas
The noise map models road traffic only, so a pedestrian street full of terraces comes out quiet: the official map puts Puerta del Sol at 36 dB. We therefore count bars, pubs and clubs within 250 metres and turn them into an estimated noise level, 40 + 10·log10(venues), added in energy to the measured traffic. It is the arithmetic an acoustician uses to add two sources: doubling the venues adds 3 dB, not double. Weeknights get 3 dB less, which amounts to assuming half the activity. The result is always labelled estimated, never measured.
Weekdays and weekends
Two scores are calculated. Venues count according to their opening hours in OpenStreetMap. Only one in ten publishes them: when missing, we assume weekend-only late opening, and the page marks it as estimated.
The sleep facts
Soundproofing, blackout, mattress, pillow and climate control come from the hotel's official website, extracted automatically and always with the literal quote behind them. If the quote is not on the site, the value is dropped. Hotels can complete or correct their own page.
What we don't do
- No platform reviews and no Google content.
- Scores are never changed for commercial reasons.
- Where there is no data, we say so.