TurkeyTR
climbing (+21.9pp)30-day day-by-day shutdown-risk trajectory. Each point is a per-horizon XGBoost prediction (h=1..30) with a 90% conformal band derived from absolute LOCO residuals. Peak risk highlighted.
Peak day
53.4%
d6 · 2026-08-19
Floor day
1.5%
d3 · 2026-08-16
30-day delta
+21.9pp
d30 vs d1 probability
30-day trajectory with 90% conformal band
model trajectory_v1mean P(shutdown by day)90% conformal bandpeak day
Drivers (SHAP top 3 from the 7-day model — same features feed all horizons)
day 115.3%
2026-08-14
- block_rate_roll7_std-0.031
- week_of_year-0.023
- gdelt_unrest_30d-0.018
day 716.4%
2026-08-20
- block_rate_roll7_std-0.031
- week_of_year-0.023
- gdelt_unrest_30d-0.018
day 1426.5%
2026-08-27
- block_rate_roll7_std-0.031
- week_of_year-0.023
- gdelt_unrest_30d-0.018
day 3037.2%
2026-09-12
- block_rate_roll7_std-0.031
- week_of_year-0.023
- gdelt_unrest_30d-0.018
Honest caveats
- Per-horizon XGBoost stacked into a trajectory — not a true joint distribution.
- Conformal bands are absolute-residual q90 from LOCO predictions on 18 eval countries.
- Long horizons (h>=14d) have wider effective uncertainty than the conformal q90 suggests; the same lag features feed all horizons.
- Bands floor at [0,1]; 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) | 90% interval |
|---|---|---|---|
| d1 | 2026-08-14 | 15.3% | [0%, 59%] |
| d7 | 2026-08-20 | 16.4% | [0%, 81%] |
| d14 | 2026-08-27 | 26.5% | [0%, 95%] |
| d30 | 2026-09-12 | 37.2% | [0%, 100%] |
Spotlight countries