How Scores Work
A transparent look at how municipality scores are calculated
1. About Our Data
Data Sources
All scores on this site are based exclusively on official Japanese government statistics. No external rankings or user reviews are used — everything is derived from quantitative calculations on government data. (Some derived indicators such as heating and cooling loads are computed from official climate statistics using formulas.)
Data comes from the following sources:
| Data Source | Indicators | Year(s) |
|---|---|---|
| e-Stat (Government Statistics Portal) | Total population, foreign resident ratio, rent, temperature, sunshine, snowfall | 2007–2023 |
| Statistics Dashboard (Ministry of Internal Affairs) | Retail stores, restaurants, clinics, income, population density, crime rate, households, kei & small special vehicles, passenger vehicles (prefecture-level) | 2020–2025 |
| Retail Price Statistics Survey (Ministry of Internal Affairs) | Consumer Price Regional Index (overall + 10 categories) | 2024 |
| National Land Info. Administrative Boundaries (N03) (Ministry of Land, Infrastructure, Transport and Tourism) | Municipality representative coordinates (lat/lon) | 2024 |
| National Land Info. Land Price Survey (L02) (Ministry of Land, Infrastructure, Transport and Tourism) | Average residential land price (¥/m²) | 2025 |
This site uses these statistical datasets and their API services, but their content is not guaranteed by the Japanese government.
2. How Scoring Works
Scoring Steps
Scores are calculated in five steps.
Step 1 — Normalization (Convert to 0–1)
To fairly compare indicators with different units (e.g., rent in yen, temperature in °C), each indicator is converted to a percentile rank across all 1,917 municipalities, scaled to 0.0–1.0.
normalized = rank of municipality / (total municipalities − 1)Example: If a municipality has the 200th cheapest rent out of 991 municipalities, its normalized value is ~0.80 (top 20% cheapest). This approach prevents extreme outliers (e.g., Tokyo, Osaka) from distorting the distribution.
Step 2 — Direction Adjustment
Different indicators have different preferred directions:
- Lower is better (rent, crime rate, snowfall, etc.): value is inverted (
1 − normalized) - Higher is better (walkability, temperature, sunshine, etc.): value used as-is
The direction toggle lets you reverse the preferred direction. For example, setting rent to “high” finds premium, affluent areas.
Step 3 — Weighted Average Score
A weighted average is computed across all active indicators based on your settings.
score = Σ(adjusted_normalized × weight) / Σ(weight)| Setting | Weight |
|---|---|
| Off | 0 (excluded) |
| Slight | 0.33 |
| Prefer | 0.66 |
| Priority | 1.0 |
Indicators with missing data are excluded from that municipality's score calculation (not treated as 0).
Step 4 — Dealbreaker Filters
Dealbreakers exclude municipalities before scoring if they fail your conditions. Comparisons use raw data values (not normalized), so you can set intuitive thresholds like "exclude areas with more than 30 cm of snow."
| Indicator | Type | Default |
|---|---|---|
| Snow Depth | Max | ≤ 30 cm |
| Min Temperature | Min | ≥ −5°C |
| Max Temperature | Max | ≤ 35°C |
| Crime Rate | Max | ≤ 5 per 1,000 |
| Population Density | Min | ≥ 500 /km² |
| Clinics | Min | ≥ 1 |
| Est. Heating Load | Max | ≤ 1,500 degree-days |
| Est. Cooling Load | Max | ≤ 1,000 degree-days |
Step 5 — Low-Score Cutoff
Municipalities with a score below 0.20 are automatically excluded as essentially unrelated to your criteria.
3. Indicator Reference
Details on each indicator
📊 National Average Comparison
Each municipality's detail page shows the national average (across approximately 1,900 municipalities) alongside every indicator value. Values better than the national average are shown in green and worse in red, based on each indicator's defined 'better direction'. Neutral indicators with no inherent direction are shown in gray.
Housing & Cost
Rent (¥/tatami)
Monthly rent per tatami (≈ 1.65 m²) for private residential units. Source: Housing and Land Survey (MIC, quinquennial, latest 2023).
⚠️ ~52% coverage: Data available for only about half of all municipalities (gaps concentrated in rural and depopulated areas).
Avg. Residential Land Price (¥/m²)
Extracted from the Ministry of Land, Infrastructure, Transport and Tourism's Prefectural Land Price Survey (National Land Numeric Information, 2025). Filtered to land use type “住宅” (residential), grouped by municipality, and averaged by survey price (¥/m²). Covers approximately 14,400 residential survey points nationwide. Higher values indicate areas with higher land and living costs.
⚠️ ~98% coverage: Data available for 1,882 municipalities (rural and mountainous areas with no residential survey points are excluded).
Daily Convenience
Walkability Score (facilities/km²)
Combined count of retail stores, restaurants, and clinics divided by habitable land area. Higher values indicate better everyday infrastructure. Source: Economic Census (MIC, 2021–2022) + Land Area Statistics.
Kei Vehicles per Household (vehicles/household)Calculated
Derived from kei & small special vehicle registrations (Statistics Dashboard) divided by number of households (Population Census).
Interpretation note: A high value in rural areas suggests heavy car dependency (poor public transport), while a low value in urban areas reflects well-developed transit. Neither direction is inherently better — it depends on lifestyle preference — so this indicator is treated as neutral (no higher_is_better direction) and shown on detail pages for reference only, not included in search scores.
Passenger Vehicles per Household (vehicles/household)Calculated
Derived from passenger vehicle registrations (Statistics Dashboard) divided by number of households (Population Census). Covers all passenger cars including kei vehicles.
Prefecture-level reference value: Municipality-level data does not exist, so the prefecture value is applied to all municipalities within that prefecture. The national average is approximately 1.06 vehicles/household (vs. official 1.016 — the gap reflects the difference in household count sources: Population Census 2020 vs. Basic Resident Register 2024). Shown on detail pages for reference only, not included in search scores.
Safety
Crime Rate (per 1,000)
Annual reported criminal offenses divided by population per 1,000. Source: National Police Agency crime statistics (prefecture-level, 2022).
⚠️ Municipality-level data was discontinued after 2008. The same prefecture-level value is applied to all municipalities within a prefecture.
International Environment
Foreign Resident Ratio (%)
Share of foreign nationals in the total population ("unknown" nationality excluded). Source: Population Census (MIC, 2020).
Population & Urbanization
Habitable Area Population Density (persons/km²)
Population per km² of habitable (non-mountainous) land. “Urban” direction: prefer areas with dense amenities. “Rural” direction: prefer quieter, nature-rich areas.
Cost of Living
Consumer Price Regional Index (National avg. = 100)
Prefecture-level price index based on the Ministry of Internal Affairs' Retail Price Statistics Survey (2024). The national average is set to 100. Covers 10 expenditure categories: food, housing, utilities, household goods, clothing, healthcare, transportation & communications, education, culture & recreation, and miscellaneous — plus an overall composite index. Each municipality's detail page displays the prefecture's composite index and the top 3 most expensive/cheapest categories.
⚠️ Prefecture-level only: The same value applies to all municipalities within a prefecture. Municipality-level CPI data is not publicly available.
Climate
Annual Avg. Temp (°C) / Sunshine Hours (hrs/yr) / Max Snow Depth (cm)
All three use prefecture-level data applied uniformly to every municipality within the prefecture. Source: Meteorological statistics (JMA via Social/Population Statistics System).
⚠️ Snow depth data is from 2007 only (confirmed via API: no data published after 2008).
Est. Heating Load / Est. Cooling Load (degree-days)
Composite indicators of heating/cooling demand (detail page and dealbreakers only). Calculated using monthly average temperatures from the nearest JMA AMeDAS station (Jan–Dec 2025) via the standard HDD/CDD method:
Heating (HDD18): Σ(month) max(0, 18 − avg_temp[m]) × days_in_month[m]
Cooling (CDD24): Σ(month) max(0, avg_temp[m] − 24) × days_in_month[m]
Base temps: heating 18°C (Energy Conservation Act / ASHRAE), cooling 24°C (JMA standard)4. Limitations & Caveats
What to keep in mind
Data Freshness
Data collection years differ by indicator (latest: 2023; oldest: 2007 snow depth). Scores are a snapshot in time. We recommend on-site verification before making a final decision to relocate.
Prefecture-Level Imputation
Climate indicators (temperature, sunshine, snowfall) and crime rate exist only at the prefecture level. The same value is assigned to every municipality within a prefecture. For example, all municipalities in Hokkaido share the same snow depth value, despite large intra-prefectural variation.
Rent Data Coverage
The Housing and Land Survey covers only ~52% of all municipalities. When rent is weighted, only municipalities with rent data appear in results.
Scores Are Relative
The absolute score value (e.g., 0.75) has no direct meaning. Interpret it as "relative fit among all municipalities for your chosen criteria."
5. Coming Soon
Planned features
Catchment Area View
A map feature showing facilities reachable within 30 minutes by car from a selected municipality. Planned using Mapbox Isochrone API combined with National Land Information facility data.
Transit Access Indicators
Modal share by commute method (car, rail, walking, etc.). Planned from Population Census commuting destination data.