Methods
How the figures on this site are measured and modelled. This page is short on purpose and grows with the site.
What a price is
Price per night of a 3-star hotel, for 2 adults staying 7 nights. The chart on the landing page shows the average over each check-in week; its two questions use the daily prices. Measured daily from hotel listings in 60 places. This is not a booking site: compare here, book wherever you like. Data through 12 Nov 2026.
On the map, the figure is the median modelled €/night for a 7-night, 2-adult stay at the baseline (3★, rating 4.0–4.5); the star and rating filters rescale every € figure by a modelled factor. The period in the masthead is a check-in range (median of its days) or one 7-night stay (the check-in day).
How to read the charts
line / band: from the observed daily medians around each day; the band is their spread, not scaled to sample size, and does not include how much a future date's price may still move before arrival · star tiers: dark violet = the selected tier, thin grey = the other tiers with enough data in that market (dash-dotted = pooled with neighbouring markets); tiers without enough data are not drawn · breakfast / free cancellation (box beside the headline): modelled extras for the selected star tier in €/night and %, information only, not added to any chart or figure · chart view (Price | Weather | All tiers above the chart): Price = the selected tier only, Weather = the daily high and rain chance of the market (a heat-strip row per weather region for Spain and a region), All tiers = all tiers and the national line · censored: not enough data for a reliable value (shown greyed) · filled (grey dashed): too few or too widely spread observations; no band
line = local linear fit through the 14 nearest observed days, by calendar day across the year break · shaded = ±1.96 × their spread around that line · grey dashed = their weighted mean instead, where they span more than 90 days or all lie on one side of the day (no band)
Weather on the map
Weather mode: the map colour is the typical daily high of the selected dates, interpolated between 300 weather zones (a continuous field, drawn on land only and fainter where no zone is near; hover it to see the interpolated value and the nearest zone). The legend shows it in eight 5 °C steps, the same in every period, so July and February compare. Texture: small dots = expected rain days a week (from 0.5), crosses = snow days a week. Labels: one per weather region, the mean of that region’s own zones (the build keeps them within 2.5 °C of each other): the high, one, two or three drops (light, some, rainy; an asterisk = snow likely; none = mostly dry) and a cloud mark with the mean cloud cover (outline under 30 %, light fill 30–60 %, dark fill over 60 %). Weather mode shows no prices: the dots are small position markers (click one to open it). The map always shows the typical year, 2011–2025 (reanalysis), never the forecast. No weather figure is pooled across climates or markets. The values are placeholders, not measurements.
Sources and attribution
- Prices: hotel rates listed publicly for the stay described above, collected daily.
- Weather: the typical weather of the selected dates, from reanalysis over 2011–2025 (ERA5, via Open-Meteo.com, CC BY 4.0; attribution wording to be confirmed). Average high is the mean daily high; rain days a week is the expected number of days with 1 mm or more, 7 times the share of such days; snow days likewise with 1 cm. Mostly dry is under 0.5 rain days a week, a rain day possible 0.5 to 1.5, some rain days 1.5 to 3, rainy 3 or more; snow likely from 0.5 snow days a week. Nothing is pooled across markets. The price index does not depend on the weather. The weather values are placeholders for now, not measurements.
- Visit counts: Cloudflare Web Analytics, which sets no cookies and stores nothing in your browser (placeholder wording, to be confirmed with the privacy page).
- Newsletter: Substack stores the address of anyone who subscribes; the page makes no request to Substack until you use the box.