Current data status

The published ranking is the Mobile Traveler Index. Most country values currently come from a legacy baseline snapshot dated December 2024. The original row-level source provenance for that snapshot is incomplete, so baseline records are labelled low confidence and are never presented as current measurements. Fixed-broadband and combined rankings are withheld until a reproducible, licensed source pipeline is available.

What the score represents

The Traveler Score is a 0–100 comparison intended to describe practical mobile-internet usability for remote work and travel. It combines download speed, latency, upload speed, stability and availability. A score is comparative, not a service guarantee for a particular address, carrier or tariff.

Download speed percentile35%

Latency percentile (inverted)25%

Upload speed percentile15%

Stability15%

Availability10%

Data modes and dates

Baseline: A dated reference value with no GCI community samples in that period. It retains the true source period rather than being copied into the current month.
Hybrid: A weighted blend of the baseline and fewer than 200 valid community samples collected in one calendar month.
Measured: Metrics calculated from at least 200 valid community samples in one calendar month.

The API exposes the mode, sample count, confidence level, data period and availability status so consumers can distinguish these states.

Community measurement

A visitor must explicitly start the test and consent to submitting results. The browser measures latency and download throughput against Cloudflare’s edge test endpoint; an upload value is accepted only when enough real upload samples complete. Interrupted or hidden-tab runs are rejected. The server validates ranges, country codes, origin, CSRF token and submission rate before saving a result.

Stored measurement fields include country code, network metrics, a coarse device/network category, test version and timestamps. Precise GPS coordinates and the raw IP address are not stored with measurement records. See the Privacy Policy for operational logs, retention and third-party processing.

Aggregation and confidence

Median values reduce the effect of extreme samples.
Country-month aggregation prevents a single recent test from redefining a national score.
Fewer than 50 samples is low confidence, 50–199 is medium, and 200 or more is high.
Rate limiting, input validation and test-version checks reduce automated manipulation.

Community traffic is self-selected and may cluster around cities, devices or networks. Even a high-confidence score is therefore an observational indicator, not a statistically representative household survey.

Geography and missing areas

The map separates geometry from index data. A territory may appear on the map without receiving a country score. We corrected deterministic joins for France, Norway, Kosovo and Taiwan, but do not assign a neighbouring country’s value to disputed or separately administered areas such as Northern Cyprus or Western Sahara. The full policy and unresolved coverage list are documented on Data Coverage & Map Geography.

Reproducibility and corrections

CSV exports and JSON endpoints provide the currently available mobile records. Any future external source must be licensed for this use, saved with a retrieval date and source identifier, and reproducible from the importer. Estimated legacy fixed-broadband importers are disabled.

Questions, correction evidence and source proposals are welcome through the contact form. Material methodology changes will increment the version shown on this page.