KASPR Datahaus← All stories

District brief

Iran, January 2026: twenty days below baseline

Window: 8–27 January 2026 · Iran · vintage 2026-07-16

National connectivity fell out of its normalcy band on 8 January 2026 and stayed out for twenty days. The national series bottomed near zero against its frozen baseline, and 163 of the 254 measured districts recorded negative episodes of their own inside the window. The pattern — near-total, simultaneous, province after province — is what a coordinated restriction looks like in measurement, though the data records the where and the how long, not the cause.

20

days below 90% of baseline

99%

deepest national shortfall

163

districts with their own episode

254

districts measured

The national record

One line, twenty days

The chart below is the national daily connectivity series against the baseline the anomaly model froze when the episode began. Freezing matters: a trailing baseline would chase the collapse downward and understate it, so depth is measured against what normal looked like before the window opened.

Connectivity against the frozen baseline, ±30 days

Daily share of measured addresses responding (solid) against the baseline the anomaly model froze when the episode began (dashed). The shaded span is the episode window.

A daily connectivity series collapsing far below its frozen baseline during the episode window, then recovering.

The subnational record

Province by province

The same window, one level down. Every province listed recorded its own negative episode overlapping the national one; several sat at effectively zero for the duration. District-level records behind each province carry the same fields.

Where it fell hardest
ProvinceFromDaysDepth
Qazvin2026-01-1612−100%
Bushehr2026-01-1413−100%
Hormozgan2026-01-1611−100%
Gilan2026-01-1413−100%
Yazd2026-01-1216−100%
Markazi2026-01-225−100%
Mazandaran2026-01-1612−100%
Ardebil2026-01-224−100%
Semnan2026-01-1711−100%
Kerman2026-01-0918−100%

Depth is the deepest daily shortfall against the frozen baseline during the episode, in per cent.

How to read this brief

An episode is a run of days outside the 99.9% prediction band of a fixed-effects model on hour-of-day and day-of-week over a 90-day trailing window. The brief records where connectivity fell, how deep and for how long. It is a measurement window, not an attribution of cause.

Independent, coarser-grained corroboration for national events is available from Cloudflare Radar’s outage centre, which observes traffic to its own network. Radar outage centre