How We Score
Destinations.
SafeTravelGal compiles safety information from official government advisories and a versioned verified dataset. This page describes exactly what we fetch, how often, what we cite, and what the methodology does not yet do.
Dataset v1.0 · Frozen 2026-05-13
By SafeTravelGal Data Team · First published 2026-05-30 · Last updated 2026-10-09
Download methodology (v1.0, PDF)Live data we ingest today
The feeds marked live are fetched every 6 hoursby an automated job and stored in our database; the rest are dated local datasets. Every country guide displays per-source last-fetched timestamps in its “Sources & last updated” block.
US State Dept
US Department of State
Levels 1-2 reviewed every 12 months; Levels 3-4 at least every 6 months; reactive updates as needed.
UK FCDO
UK Foreign, Commonwealth & Development Office
Updates can occur several times a day during a crisis; daily upstream feed updates.
AU DFAT
Australian Department of Foreign Affairs and Trade (Smartraveller)
Event-driven; daily upstream feed.
CDC
US Centers for Disease Control — Travel Health Notices
Issued event-driven; supplemental health alerts.
GPI
Institute for Economics & Peace — Global Peace Index
Annual dataset published by IEP. Local copy updated when each year's report releases.
Equaldex
Equaldex LGBTQ+ Rights Index
Community-maintained legal status database. Our values come from a local dataset; they are not fetched from Equaldex on the cron.
UNODC
UNODC intentional-homicide rate (via World Bank)
UNODC publishes annually; the World Bank indicator updates as national statistics are released.
CDC travel health notices are supplemental — they are not part of the composite safety score, but new notices for tracked countries are recorded in that country guide's advisory history.
Dataset v1.0: how the composite score is currently produced
Static composite, expert-reviewed at dataset freeze date. Composite overallScore is a single number on a 0 to 10 scale (for example 9.1 or 5.7) assigned by the dataset author with reference to the sources listed on the methodology page; it is not a live computation. v1.1 will introduce a live composite — see methodology.
Honest disclosure
The composite overallScore on each country guide and the four sub-scores (theft, harassment, scam, night safety) are fixed estimates, drafted with AI assistance in May 2026. We cannot point to a source for each number, and no named person has signed them off. That is why every guide shows the US and UK government advisories beside them, and why v1.1 replaces these estimates with a published formula fed by government data. Where a score has not yet been set, the field is marked as pending rather than displaying an unverified value.
v1.1 will introduce a live composite with documented weights — see Section 3.
Reference sources for the v1.0 dataset
These are the reference sources for the v1.0 dataset. They are not currently live-ingested, and we cannot show that each one shaped every v1.0 number. They are listed here for transparency. Live ingestion of each is planned for v1.1 dataset releases.
Georgetown Women, Peace, and Security Index
https://giwps.georgetown.edu/the-index/→WEF Global Gender Gap Index
https://www.weforum.org/reports/global-gender-gap-report/→Live Ingestion Pipeline Map
1. API Ingestion
Polls US State Dept, UK FCDO, and CDC RSS/JSON feeds in parallel.
2. Change Diffing
Compares current advisory scores with stored rows to identify changes.
3. Supabase Write
Updates countries table via service-role client and logs events.
4. ISR Regeneration
Triggers Vercel revalidation paths so affected country guides rebuild on their next request.
5. Search Submit
Submits changed URLs to IndexNow.
Dataset version log
v1.0 — current — frozen 2026-05-13
Static composite, 50 countries. Live updates to advisory level and FCDO summary apply per-row as the cron fires; the composite overall score does not move until v1.1.
v1.1 composite formula — published, not what sets today's scores
Equally weighted average of four normalized safety indicators. Each source is independently normalized to a 0–100 scale (100 = safest). The composite is the arithmetic mean of available sources, then scaled to 0–10 for display. Minimum 2 sources required; otherwise the score is omitted.
A limit of the UK FCDO row: it scores keywords in the stored FCDO text. In the v1.0 dataset file that text is GOV.UK's generic page description for 48 of the 50 countries, which scores 75 for every one of them, so for those 48 this row does not separate one country from another (checked against the dataset file on 2026-10-09). The other two carry a placeholder that scores 100, which the FCDO did not write. The live site now stores the FCDO's own update notes instead, and 48 of the 50 match no keyword (read from the live table on 2026-10-09), so they also score the default 75.
Output: 0–100 normalized → 0–10 displayed. Minimum 2 sources required.
Reproducibility
Every input, weight, normalization rule, and source URL of the v1.1 formula above is published on this page, so a reader can work out what that formula gives for any country. That is not the same as reproducing the v1.0 scores shown on country guides. Those are fixed estimates (see the honest disclosure above), and they are not the output of this formula. Run on the dataset file on 2026-10-09, the formula reproduces 47 of the 50 composites; three were set lower by hand (Czech Republic 9.1 against 9.7, United Kingdom 8.2 against 8.7, Israel 5.9 against 6.1). The scores the guides display come from the live database and are whole numbers, which this formula does not produce. When the composite moves to live computation in v1.1, every variable change will be logged in the public methodology changelog.
Change the weights, see the ranking move
The v1.1 composite weights the four indicators equally. Drag a slider to weight one more heavily and every country re-ranks live. The inputs below are the real normalized values for 50 countries, produced by applying the normalization rules published above to this site's own dataset — not example figures.
What these numbers are, and are not
- These are v1.1-computed scores. The score on a country's own guide is v1.0, and the two will differ. The published v1.0 dataset is expert-reviewed and frozen; this sandbox applies the computed formula to live inputs. A gap between them is the formula changing, not a country changing.
- A country is ranked only when at least 2 of the four indicators have a value for it, matching the real pipeline. Missing sources are dropped and the weights renormalized across what is left, so an absent feed never scores as zero.
- The AU DFAT values are dataset values captured at dataset freeze, not a live feed. Smartraveller blocks our automated requests, which is stated in full in the sources section above. The indicator stays in the formula because the formula is published; the values behind it are dated.
Ranked output
v1.1 computed50 of 50 countries have enough sources to score under the current weighting.
- #1Czech Republic3/4 sources9.6/10
- #2Iceland3/4 sources9.0/10
- #3Ireland3/4 sources8.9/10
- #4Austria3/4 sources8.8/10
- #5New Zealand3/4 sources8.8/10
- #6Portugal3/4 sources8.8/10
- #7Singapore3/4 sources8.8/10
- #8Switzerland3/4 sources8.8/10
- #9Australia3/4 sources8.7/10
- #10Canada3/4 sources8.7/10
- #11Croatia3/4 sources8.7/10
- #12Finland3/4 sources8.7/10
- #13Japan3/4 sources8.7/10
- #14Norway3/4 sources8.7/10
- #15Malaysia3/4 sources8.6/10
- #16Taiwan3/4 sources8.6/10
- #17Greece3/4 sources8.5/10
- #18South Korea3/4 sources8.5/10
- #19Vietnam3/4 sources8.5/10
- #20Argentina3/4 sources8.4/10
- #21Thailand3/4 sources8.1/10
- #22Denmark3/4 sources7.6/10
- #23Germany3/4 sources7.5/10
- #24Netherlands3/4 sources7.5/10
- #25Spain3/4 sources7.5/10
- #26Sweden3/4 sources7.5/10
- #27Costa Rica3/4 sources7.4/10
- #28Italy3/4 sources7.4/10
- #29United Kingdom3/4 sources7.4/10
- #30Cambodia3/4 sources7.2/10
- #31France3/4 sources7.2/10
- #32Indonesia3/4 sources7.2/10
- #33Nepal3/4 sources7.2/10
- #34Sri Lanka3/4 sources7.2/10
- #35Ecuador3/4 sources7.0/10
- #36Peru3/4 sources7.0/10
- #37Morocco3/4 sources6.9/10
- #38Tanzania3/4 sources6.9/10
- #39Brazil3/4 sources6.8/10
- #40India3/4 sources6.8/10
- #41Philippines3/4 sources6.8/10
- #42Egypt3/4 sources6.7/10
- #43South Africa3/4 sources6.7/10
- #44Mexico3/4 sources6.5/10
- #45Turkey3/4 sources6.5/10
- #46Jordan3/4 sources6.2/10
- #47Guatemala3/4 sources5.7/10
- #48Kenya3/4 sources5.7/10
- #49Colombia3/4 sources5.4/10
- #50Israel3/4 sources4.8/10
Source freshness
Pull cadence and last check for every data source used by SafeTravelGal. Live government advisory feeds are scheduled on a 6-hour cron and show the date of their last ingest run, read from our ingest record; datasets show the date they were captured or frozen.
| Source | Canonical URL | Pull cadence | Last checked |
|---|---|---|---|
| US State Dept | https://travel.state.gov/content/travel/en/traveladvisories/traveladvisories.html | every 6 hours | Last run 2026-10-20 |
| UK Foreign, Commonwealth & Development Office | https://www.gov.uk/foreign-travel-advice | every 6 hours | Last run 2026-10-20 |
| Australian Department of Foreign Affairs and Trade (Smartraveller) | https://www.smartraveller.gov.au/ | not ingesting (dataset v1.0, captured 2026-05-26) | Not checked: blocked since 2026-05-17 |
| US Centers for Disease Control — Travel Health Notices | https://www.cdc.gov/travel-notices/ | every 6 hours | Last run 2026-10-20 |
| Institute for Economics & Peace — Global Peace Index | https://www.economicsandpeace.org/global-peace-index/ | annual (local dataset) | Local dataset, frozen 2026-05-13 |
| Equaldex LGBTQ+ Rights Index | https://www.equaldex.com/ | local dataset (not a live pull) | Local dataset, not a live pull |
| EmergencyNumberAPI | https://emergencynumberapi.com | At dataset generation | Seeded at dataset freeze, 2026-05-13 |
About the dataset: SafeTravelGal Solo Female Safety Index v1.0
Safety scores and risk sub-scores for solo female travelers across 50 countries (dataset v1.0). The US State Dept and UK FCDO advisory fields and CDC health notices are checked separately, on a schedule of every 6 hours.
Version v1.0. Frozen 2026-05-13. Covers 2026. Coverage: Worldwide (50 countries).
What it measures
- US State Department advisory level (live)
- UK FCDO travel advice update note (live)
- AU DFAT advisory level (dataset v1.0, captured 2026-05-26; not live)
- CDC travel health notice (live, supplemental)
- Global Peace Index score and rank (annual)
- Equaldex LGBTQ+ legal status
- Theft sub-score (v1.0 estimate)
- Harassment sub-score (v1.0 estimate)
- Scam risk (v1.0 estimate)
- Night safety (v1.0 estimate)
- Emergency service numbers (dataset v1.0)
- Composite overall safety score (v1.0 estimate)
What this methodology does NOT yet do
Not Ingested
Georgetown WPS Index, WEF Gender Gap. These remain reference sources for the v1.0 dataset, not live feeds. Country pages will not claim them as live data. The UNODC homicide rate is ingested live (see Section 1). The Global Peace Index and Equaldex are read from local datasets, not fetched live.
Not Computed
At v1.0 the composite overall score is a static value a fixed estimate set at dataset freeze (2026-05-13) and drafted with AI assistance, with no source recorded for each number. It is not a live algorithmic computation and not the output of the formula in Section 3. The live composite formula in Section 3 ships with v1.1, where countries with fewer than 2 available sources will show no score rather than an unreliable estimate.
Not Absolute
No safety score is a guarantee. Conditions change rapidly; always check your government's current advisory before booking or traveling. SafeTravelGal content is informational only, not a substitute for current official guidance or licensed professional advice.
Not First-Person
First-person travel anecdotes about a destination only appear when the named verifier or a named contributor has personally visited that destination. Desk-researched guides are labeled explicitly and cite the sources used.
Data verification
No Data Verification Lead has been named yet. Country, ranking, gear, and insurance pages are published under the SafeTravelGal Data Team byline, not under a named person; see /about for status.