voidly

KazakhstanKZ

climbing (+9.8pp)

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
55.8%
d14 · 2026-08-27
Floor day
1.0%
d3 · 2026-08-16
30-day delta
+9.8pp
d30 vs d1 probability

30-day trajectory with 90% conformal band

model trajectory_v1
0%25%50%75%100%d1d7d14d21d30peak d14 56%
mean P(shutdown by day)90% conformal bandpeak day

Drivers (SHAP top 3 from the 7-day model — same features feed all horizons)

day 19.7%
2026-08-14
  • block_rate_lag1-0.029
  • month-0.024
  • week_of_year-0.023
day 714.1%
2026-08-20
  • block_rate_lag1-0.029
  • month-0.024
  • week_of_year-0.023
day 1455.8%
2026-08-27
  • block_rate_lag1-0.029
  • month-0.024
  • week_of_year-0.023
day 3019.5%
2026-09-12
  • block_rate_lag1-0.029
  • month-0.024
  • week_of_year-0.023

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

DayDateP(shutdown)90% interval
d12026-08-149.7%[0%, 54%]
d72026-08-2014.1%[0%, 79%]
d142026-08-2755.8%[0%, 100%]
d302026-09-1219.5%[0%, 98%]
Spotlight countries