Carregando…
Foz do Iguaçu · Ciudad del Este · Puerto IguazúSaturday, July 18, 2026
Crowd probability · in-house model
Direct answer: Today the chance of a packed house is high at: Iguaçu Falls (BR), Itaipu, Bird Park, Three Borders Landmark. Next day with all attractions quieter: Sun 02 Aug.
For each of the next 28 days, the Marco Três model computes the probability of a BUSY DAY at each attraction — crossing the real seasonality of 390 months of official data with day of week and public holidays from Brazil, Argentina and Paraguay. Probability, never a guarantee.
Next 28 days · chance of a busy day
| day | Iguaçu Falls (BR) | Itaipu | Bird Park | Three Borders Landmark | holiday |
|---|---|---|---|---|---|
| Sat 18 Jultoday | 86% | 86% | 86% | 86% | |
| Sun 19 Jul | 86% | 86% | 86% | 86% | |
| Mon 20 Jul | 29% | 29% | 86% | 86% | |
| Tue 21 Jul | 29% | 29% | 71% | 57% | |
| Wed 22 Jul | 43% | 29% | 86% | 86% | |
| Thu 23 Jul | 86% | 71% | 86% | 86% | |
| Fri 24 Jul | 86% | 86% | 86% | 86% | |
| Sat 25 Jul | 86% | 86% | 86% | 86% | |
| Sun 26 Jul | 86% | 86% | 86% | 86% | |
| Mon 27 Jul | 29% | 29% | 86% | 86% | |
| Tue 28 Jul | 29% | 29% | 71% | 57% | |
| Wed 29 Jul | 43% | 29% | 86% | 86% | |
| Thu 30 Jul | 86% | 71% | 86% | 86% | |
| Fri 31 Jul | 86% | 86% | 86% | 86% | |
| Sat 01 Aug | 57% | 29% | 57% | 43% | |
| Sun 02 Aug | 14% | 14% | 14% | 14% | |
| Mon 03 Aug | 14% | 14% | 14% | 14% | |
| Tue 04 Aug | 14% | 14% | 14% | 14% | |
| Wed 05 Aug | 14% | 14% | 14% | 14% | |
| Thu 06 Aug | 14% | 14% | 14% | 14% | |
| Fri 07 Aug | 14% | 14% | 14% | 14% | |
| Sat 08 Aug | 57% | 29% | 57% | 43% | |
| Sun 09 Aug | 14% | 14% | 14% | 14% | |
| Mon 10 Aug | 14% | 14% | 14% | 14% | |
| Tue 11 Aug | 14% | 14% | 14% | 14% | |
| Wed 12 Aug | 14% | 14% | 14% | 14% | |
| Thu 13 Aug | 14% | 14% | 14% | 14% | |
| Fri 14 Aug | 14% | 14% | 14% | 14% |
How to read: a "busy day" = a day among that attraction's top 25% busiest of the year (cut at the 75th percentile of its own historical distribution). The % is a frequency: in how many of the observed years (2019, 2022–2025) that day would land above the cut. By construction it is never 0% or 100%.
Verified accuracy
Walk-forward validation: the model learns the seasonal shape only from PRIOR years and predicts high/low month by month for the following year (2022, 2023, 2024, 2025) — the test year never enters the fit. Result: 91% accuracy (175 of 192 attraction-months).
Holidays in the window
No public holiday in the 3 countries over the next 28 days.
Method & backing
The model's spine = REAL per-attraction seasonality (390 months of official SMTU Foz/DVEET data; scenario years: 2019, 2022, 2023, 2024, 2025, excluding 2020–21/pandemic). Each year is a scenario; day of week and holidays (Nager.Date, BR+AR+PY, incl. bridge days) multiply the month index; school holidays already live inside July's and January's real weight.
Honesty: the day-of-week factor is a documented heuristic — no public daily series per attraction exists (that is what a turnstile partnership would unlock). The model predicts crowd patterns, not exact counts.
Sources (full credit)
Go deeper
When to go (month by month) →Hourly flow (today & 7 days) →Queue-free itinerary →Forecast accuracy →