214 countries and territories189 of 193 UN membersmonthly to 2026-09back to 1926
Nine statistics worked out every night from each country's own official figures: seasonally adjusted trend, recovery against 2019, regional benchmarks, length of stay, 24-month forecasts with measured error, nowcasts of unpublished months, anomaly detection, origin-market shifts and market concentration. Every number comes with its method, its range and its evidence.
Read the pattern underneath the noise: what the trend really is, how far a market has come back, how it does against its neighbours, and how long visitors stay.
Look ahead and catch the turn early: where a country is heading, what the months not yet published will show, which months broke the pattern, which markets are rising, and how exposed a destination is.
What each statistic returns, field by field. Nothing is summarised away: the method, the window and the inputs come back with every answer.
| Statistic | Endpoint | Fields it returns | Parameters | Method in meta | Countries | Plan |
|---|---|---|---|---|---|---|
| Trend | /v1/stats/trend | period, adjusted, trend | country, from | seasonal_factors, method, series | 77 | Analyst |
| Recovery | /v1/stats/recovery | period, index | country, from | basis, series, first, last | 84 | Analyst |
| Benchmark | /v1/stats/benchmark | year, country, region | country, peers | region, peers, basis | 72 | Analyst |
| Stay length | /v1/stats/stay-length | period, nights_per_arrival | country, from | series, unit | 2 | Analyst |
| Forecast | /v1/stats/forecast | period, value, low, high | country, horizon, level | backtest_mape_pct, fitted_from, horizon_months, made, method | 73 | Intel |
| Nowcast | /v1/stats/nowcast | period, value, low, high, estimated | country, level | published_to, estimated_to, yoy_used, method | 63 | Intel |
| Anomalies | /v1/stats/anomalies | period, score, expected | country, min_score | months_checked, flagged, threshold_sd | 77 | Intel |
| Origin shift | /v1/stats/origin-shift | origin, share_pct, change_points | country, limit | window, against, value | 18 | Intel |
| Concentration | /v1/stats/concentration | hhi, top3_share_pct, top, markets | country | window, scale | 18 | Intel |
A statistic is given only where the history supports it. Gold: carried now. Grey: not yet.
Every plan reads the same figures. The higher plans add the work: history, revisions, and the statistics that would take an analyst weeks to build and maintain.
The same nine statistics answer very different questions.
Plan rates and staffing a season ahead from each market's forecast.
/v1/stats/forecast?country=THSpot which markets are rising before the schedules are set.
/v1/stats/origin-shift?country=JPReport recovery against 2019 and benchmark against the region.
/v1/stats/recovery?country=CLWatch the months not yet published and the ones that broke pattern.
/v1/stats/nowcast?country=ESHow every statistic is made
estimated: true and a range on every nowcast and forecastmeta.method names the model, window and inputsOne call, the whole picture
# where a country is heading, with its error curl "https://api.tourismapi.com/v1/stats/forecast?country=AG" \ -H "X-API-Key: $TOURISM_API_KEY"
Intel includes everything in Analyst and Developer: all six measures, full history, revisions, bulk files and 1,000,000 calls a month. Cancel any time.