Turkey TR
climbing (+15.0pp)Follow the model’s 30-day path. Read the probability and uncertainty together.
30-day model trajectory
model trajectory_v1mean P(shutdown by day)Uncertainty interval withheld: 90% coverage has not been validated.peak day
Peak day
41.7%
d19 · 2026-12-07
Floor day
2.8%
d7 · 2026-11-25
30-day delta
+15.0pp
d30 vs d1 probability
Drivers (SHAP top 3 from the 7-day model — same features feed all horizons)
day 15.3%
2026-11-19
- week_of_year-0.246
- blocked_count_roll14_mean-0.050
- block_rate_roll30_mean+0.047
day 72.8%
2026-11-25
- week_of_year-0.246
- blocked_count_roll14_mean-0.050
- block_rate_roll30_mean+0.047
day 1437.3%
2026-12-02
- week_of_year-0.246
- blocked_count_roll14_mean-0.050
- block_rate_roll30_mean+0.047
day 3020.3%
2026-12-18
- week_of_year-0.246
- blocked_count_roll14_mean-0.050
- block_rate_roll30_mean+0.047
Honest caveats
- Per-horizon XGBoost stacked into a trajectory — not a true joint distribution.
- The stored LOCO absolute-residual widths are not validated 90% intervals for these forecast probabilities; intervals are withheld.
- Long horizons (h>=14d) have additional uncertainty because the same lag features feed all horizons.
- Probabilities below 0.05 are at the model's noise floor.
- SHAP attribution above is from the 7-day point model. The trajectory uses 30 separate per-horizon XGBoost models that share the same feature set; we have not yet computed per-horizon SHAP.
Sample horizons
| Day | Date | P(shutdown) | Interval status |
|---|---|---|---|
| d1 | 2026-11-19 | 5.3% | Unvalidated |
| d7 | 2026-11-25 | 2.8% | Unvalidated |
| d14 | 2026-12-02 | 37.3% | Unvalidated |
| d30 | 2026-12-18 | 20.3% | Unvalidated |
Spotlight countries
About this view
30-day day-by-day shutdown-risk trajectory. Each point is a per-horizon XGBoost model score (h=1..30). The stored LOCO residual widths are not validated probability intervals, so no uncertainty band is shown. Peak score highlighted.