Methodology and sources

Why there is no score yet

A composite publishes only when at least 60% of index weight is sourced to official data or independent measurement. Today it is 24.7% — 4 of 15 indicators. Estimates and press-reported figures deliberately do not count toward that threshold, because a score resting on Umran's own guesses is the failure the Recovery Index correction taught us to design against. Until the threshold is met, there is no composite to render: the function that would produce it returns nothing.

Why the gate will stay shut for a while

The arithmetic is not close, and it is worth being plain about why. Network quality carries a fifth of the index and currently stands at zero of three indicators. The only sources that would qualify — IODA, Cloudflare Radar, M-Lab — are not documents someone can go and read; they are interfaces returning series that change daily. A number copied out of one today is wrong next week, which is the opposite of what a measured indicator is for. Filling those rows means a scheduled fetch writing into this dataset on a timer, and until that exists no amount of research moves the gate. Coverage and price are worse: they need disclosure by operators and the regulator that does not currently happen at all, so they may never qualify.

One caveat already built in

When the speed indicator does arrive it will not be a national average, because it cannot honestly be one. M-Lab's tests are user-initiated — people run a speed test when their connection is misbehaving — so volume rises during outages, and under rolling power cuts that skew is severe. It will be published as the median throughput among users who ran a test, with the number of tests beside it, which is M-Lab's own guidance. If the Syrian sample turns out to be hundreds of tests rather than thousands, the indicator will not be published at all and the network-quality domain will be re-scoped to reachability, which IODA and Radar can measure properly. That change of weights would itself be a methodology correction, logged as one.

The formula

Every indicator is scored as progress from where Syria actually stood on 8 December 2024 toward a stated target. Identical baseline and formula to the site-wide Recovery Index, so the two can be read together.

progress % = (current − Dec 2024 baseline) ÷ (target − baseline)

Relationship to the Recovery Index

The national Recovery Index does not currently contain a telecom indicator. Its infrastructure domain covers electricity hours, generation capacity, water, health, education and transport — telecom is absent. So this section does not claim to reconcile with the national composite, and any statement that it decomposes the infrastructure domain would be false today. Adding two or three telecom indicators to the national index would change a published figure, and that is an editorial act requiring a corrections entry — not something to slip into a data file.

Recovery Index

Domains and weights

Unweighted mean of the indicators within a domain, weighted sum across domains — so a domain does not gain influence by being subdivided more finely. The seven weights sum to 1.

• Coverage 1 of 3

  Who can reach a network at all — the precondition for every other domain.

• Network quality 0 of 3

  Whether the connection works when it is needed. The domain most open to independent measurement, and so the likeliest route to opening the gate.

• Infrastructure 2 of 5

  Backbone, fibre and data centres actually in the ground — not announced.

• Competition 1 of 1

  How many operators can actually contest the market, and on what terms.

• Affordability 0 of 1

  What connectivity costs against what a household earns.

• International connectivity 0 of 1

  Capacity and redundancy out of the country — a single-path network is one cut from dark.

• Digital readiness 0 of 1

  What the network is used for once it exists: services, payments, government online.

Confidence tiers

Six tiers in the data, three badges for the reader. The tier is always visible in a figure's drill-down and in the page's source register.

Official 53

Stated by the body with authority over the fact: a ministry, the regulator, an operator about its own commitments, or a signed agreement. A press release is official for its author's own promises and reported for everything else in it.

Independent measurement 0

Measured, not reported — IODA, Cloudflare Radar, M-Lab, RIPE, BGP. These need nobody's cooperation, which is why they are the only realistic route to publishing a composite.

Reported 10

Carried by credible trade press or a secondary outlet that is not itself the authority for the claim. Useful, quotable, and not strong enough to build a score on.

Derived 0

Calculated by Umran from two or more sourced values. The working is published with it.

Umran estimate 0

Umran's own estimate, with the method disclosed. Never used to fill a gap quietly.

Not available 0

Established as unobtainable. Rendered, not hidden.

Not available 10

Not yet researched. Renders identically to unavailable, because the reader is owed the same answer either way.

Missing values

A figure Umran cannot source is null in the data and “No data” on the page. It is never zero, never interpolated, never carried forward from an earlier period, and never quietly replaced by a similar-looking number from a different definition. The build fails if a record without a source carries a value.

Commitments

The commitments ledger is documentary. Statuses are date facts recomputed at build time from each deadline, so a stored status cannot go stale. Where two claims cannot both hold, both are published and neither is resolved — averaging them or preferring the more recent would manufacture a fact.

Commitments

The map

Governorate resolution only. Umran publishes no coordinates for towers, exchanges, gateways or any other live network facility, and the data schema has no field that could hold them. Governorate boundaries are administrative and the occupied Golan renders as Syrian territory.

See every indicator