For each of 90 regions across Spain (17), Italy (20), Portugal (6), Greece (13), France (13), Croatia (21), Caterelo collects 13 signals. Seven of them are weighted into the LifeTrend™ score (§02b); the other six are published as context and used by Match Score, Hidden Gem detection and the Decision Matrix:
| Dimension | Signals | Sources |
|---|---|---|
| Property & Investment | €/m² trends 1Y/5Y/10Y (trajectory modelled), rental yield, climate trajectory 2050 | National stat offices, Eurostat |
| Cost & Income | Cost of living (single + family), avg net salary, purchasing power | INE, ISTAT, INSEE, ELSTAT, DZS, Numbeo cross-check |
| Health & Safety | Healthcare quality + cost, regional safety index, climate risk | WHO, OECD, Numbeo, Eurostat crime stats, IPCC AR6 |
| Lifestyle & Work | Coworking density, English proficiency, internet speed, sunny days | Speedtest, EF EPI, Coworker.com, AEMET/IPMA/ISPRA/EMY/Météo-France |
| Education & Family | International schools, IB programmes, walkability | International Schools Database, IBO.org |
Caterelo provides two distinct scores per region:
The core insight: there is no single objectively best region. A region is good or bad depending on the user's situation, budget, priorities and constraints. LifeTrend tells you the regional baseline; Match Score tells you whether it fits your specific life.
The Decision Matrix view (in-app) compares your shortlisted regions side by side. Decision Matrix highlights the most decision-relevant signals from the full 13-signal model, sorted by best fit. No spreadsheet needed.
For full transparency, here is every signal Caterelo computes per region — numbered, sourced, and tagged by tier (see §03 below for tier definitions).
Seven of these thirteen carry weight in the LifeTrend™ score — 02 cost of living, 05 climate (current), 07 safety, 08 healthcare, 09 digital infrastructure, 10 lifestyle, 11 education. The remaining six — 01 property price, 03 earnings, 04 rental yield, 06 climate 2050, 12 Eurostat NUTS2 baseline, 13 search interest — are published as context and used by Match Score, Hidden Gem detection and the Decision Matrix. They do not move the LifeTrend number.
Every region gets a single LifeTrend score (0-100) which weighs seven dimensions to produce a unified relocation index. The weights below sum to exactly 100%. They are explicit, not hidden:
After the raw weighted sum, the score is min-max normalized across all 90 regions and stretched to a display range of [30, 90] for human readability. The min and max are computed from the loaded dataset rather than fixed — across the current 90 regions the raw range is [34.4, 72.4]. (A hard-coded [15, 78] fallback exists in the code but applies only if no region data loads at all.) Final score range: 30-90 (mean 56.5, stdev 12.1).
Caterelo's data sources are categorized into three tiers based on authority, reproducibility, and verifiability. Every source is named on this page — 38 in total, plus a handful of collective sources we cannot enumerate individually (national health ministry reports, national seismic hazard authorities, 2021-round censuses, official pilgrimage and cultural route bodies):
Every signal in §02a above is tagged with its tier. Region pages and comparison pages display a Data Confidence badge (see §03b) reflecting actual coverage of these tiered sources for that region.
Price benchmarks (€/m²) are calibrated to national statistical offices. An automated cross-check against market portals is built but not enabled in production — instead we ship portal links on every region page so you can check current asking prices yourself. Historical price trajectories (5Y/10Y change, CAGR, momentum) are modelled from the calibrated base using a deterministic growth+volatility function — directional estimates for comparison, not measured quarter-by-quarter transaction data. Measured national house price indices (Eurostat prc_hpi_q) are shown alongside them.
score = 100 × (0.35·pop + 0.35·occupancy + 0.15·routes + 0.15·rituals), where pop = clamp((10-year population trend % + 10)/20, 0, 1), occupancy = clamp((45 − non-primary-dwelling %)/35, 0, 1), routes = min(official pilgrimage/cultural routes, 3)/3, rituals = min(region-rooted UNESCO intangible-heritage entries, 3)/3; score is null where population or occupancy data is unavailable. Sources: Eurostat/national statistics offices (2015–2025 population series), 2021-round censuses (non-primary = second/seasonal homes + vacant dwellings), ich.unesco.org, official route bodies; adversarially verified, compiled 2026-07Every region page and comparison page on Caterelo displays a Data Confidence badge (High / Medium / Low) reflecting actual data coverage for that region. The label is computed at build-time from real data presence — not estimated, not editorial.
For region pages, confidence is calculated from 22 sub-fields across the 11 data layers (cost has 3 sub-fields: single, family, rent; safety has 3: score, crime, night safety; climate has 3; healthcare 2; digital 2; lifestyle 2; education 1; rental 2; climate projection 2; earnings 1; property price 1).
| Label | Coverage | What it means |
|---|---|---|
| High | ≥ 90% | Region has full data across all 22 sub-fields. Confident in cross-region comparison. |
| Medium | 70-89% | Some data gaps — usually in newer regions or thin sub-fields. Verdict is directionally correct but verify specific gaps before deciding. |
| Low | < 70% | Material data gaps. Use for orientation only; consult original sources directly. |
For comparison pages, confidence reflects whether both compared regions have data in each of the 6 winner-by-category dimensions. A 100% Confidence comparison means every dimension has data on both sides; lower scores indicate one or both sides have gaps.
There is no automated full refresh. Different layers move in different ways, and we would rather state that than publish a cadence we do not keep:
prc_hpi_q) and ECB mortgage rates (MIR cost-of-borrowing). Pulled by scripts/fetch-eurostat-hpi.js and scripts/fetch-ecb-rates.js. Last run: 3 July 2026 (latest published quarter at that point: 2026-Q1).Last full review of the base data: Q4 2025 (property price base and regional benchmarks). Layers added or revised since then are listed below. We are not publishing a next-refresh date until the refresh is actually scheduled.
prc_hpi_q) alongside modelled regional trajectories; data vintage remains Q4 2025 for base prices.Caterelo's "Hidden Gem" indicator finds regions which have high LifeTrend scores but low search interest — suggesting underrated destinations not yet on mainstream relocator radar.
Formula: gem_ratio = LifeTrend ÷ (Google_search_interest + 1). Higher ratio = more underrated. Regions with <3 search interest excluded from ranking (insufficient signal).
Forward-looking climate projections use IPCC AR6 SSP2-4.5 moderate scenario with these per-region fields:
Sources: IPCC AR6 WG1 (2021), EURO-CORDEX regional downscaling, JRC PESETA IV. This is forward-looking modeling, not certainty.
LifeTrend is a weighted composite of seven dimensions: Safety (22%), Climate (18%), Cost of Living (18%), Healthcare (13%), Lifestyle (10%), Digital infrastructure (10%), Education (9%). Each dimension normalizes raw signals to 0-1, multiplies by its weight, sums to a raw score, then min-max normalizes across all 90 regions and stretches to a 30-90 display range. Same formula across all regions — no per-region tuning. See §02b for the full breakdown.
LifeTrend is the global region score — same for every user. Match Score is personalized based on your editable Relocation Profile (situation, budget, priorities, constraints). A region with LifeTrend 71/100 may show Match 86% for a remote family seeking sun + value, but Match 52% for someone prioritizing cool summers. See §02.
There is no automated full refresh. Two layers are pulled by script on demand — the Eurostat house price index (prc_hpi_q) and ECB mortgage rates — last run 3 July 2026. Two layers are live API proxies with a 24-hour edge cache: Eurostat NUTS2 (population, income, unemployment) and Google Trends regional demand. Every other dataset, including the IPCC AR6 climate projections, is compiled and updated manually. Last full review of the base data: Q4 2025. See §04.
Data Confidence (High / Medium / Low) reflects actual data coverage for that specific region across 22 sub-fields in 11 data layers. High = ≥90% coverage, Medium = 70-89%, Low = <70%. Computed at build-time from real data presence — not editorial. See §03b.
Sources are categorized into three tiers: Tier 1 (official statistics — Eurostat, INE, ISTAT, INSEE, ELSTAT, INE PT, DZS, IPCC AR6, WHO, OECD, national meteorology services); Tier 2 (third-party indices with transparent methodology — Numbeo, Speedtest, EF EPI, property portals); Tier 3 (derived signals computed inside Caterelo — Google Trends, Hidden Gem ratio, Match Score, composite LifeTrend). See §03.
No. Caterelo is a decision-support tool that provides informational data and comparative signals for relocation research. It is not legal, tax, immigration, real estate or investment advice. Always verify country-specific regulations and consult licensed professionals before making relocation, property or financial decisions.
Climate projections use IPCC AR6 SSP2-4.5 (moderate emissions scenario) with EURO-CORDEX 12.5km regional downscaling and JRC PESETA IV impact assessment. Per-region fields include warming by 2040, warming by 2050, summer warming, additional heat days, drought risk, sea level risk, wildfire risk. This is forward-looking modelling, not certainty. See §06.
Most relocation tools score regions through opaque algorithms. Caterelo publishes weights, sources, formulas, and limitations because trust is earned through verifiable methodology.
If you find an error or have data we should incorporate, contact us. Methodology is iterated based on real feedback from users + advisors.
SoftwareApplication + Dataset. Target audience: remote workers, founders, and families considering relocation to Spain, Italy, Portugal, Greece, France, Croatia. Free tier available; premium Decision Pack €79 one-time (180-day access). Methodology last updated: 2026-05-05.