Loading data for JapanLining up the latest figures for all 47 prefecturesPulling together quakes, weather, and live camerasBuilding the map layers

Methodology — sources and definitions

84 metrics — every source and definition, on one page

This is the place to open the moment you think "is this number really right?". The source URL, what is counted, the update cycle, and whether higher or lower is better are gathered here for 46 agencies and 84 metrics. In the world of audit, how a number was made is a stronger basis for a decision than the number itself. I laid out the metrics in the same spirit.

Operating principles

No hiding behind a number's name

A composite metric that sums up "ease of raising children" in one number always bakes the maker's convenience into the calculation. So we keep one metric to one concept and lay out childcare waitlists, daycare capacity, average birth rate, and child allowances independently. We preserve, at the layout stage, a state in which readers can interpret with their own weighting.

"Has money" and "is generous" are different stories

A high Fiscal Capacity Index (the potential to cover administrative costs from tax revenue) and a short childcare waitlist (a service actually reaching people) are related but separate. Mix them into one ranking and the counter-intuitive fact — "has money yet a long waitlist" — vanishes. So we keep them apart. We do not erase the tensions; we show them.

Not a tool for investment decisions

What is laid out here is a secondary visualisation of public statistics; it should not be used for individual property transactions or valuing individual companies. Statistics never replace "walking the ground" or "asking a professional". We design so as not to reproduce the failure I saw so often on audit engagements: deciding on numbers alone and stripping out the feel of the ground.

We do not borrow the authority of public statistics

If our figures ever disagree with the source, the source is always right. This is a layer of visualisation and interpretation; it does not stand above the original statistics. When you cite, write the original URL, not "according to Nippon Pulse". There is no value in us standing out.

Metric catalogue (84 metrics · 9 categories)

Economy(21)

Metric	Unit	Source	Update	Reading	Tier
Unemployment rate Share of people who want to work but cannot find a job. Lower is better.	%	MIC — Labour Force Survey	Quarterly	Lower is better	T1
Income per resident Income per resident, estimated from gross prefectural product.	¥	Cabinet Office — Prefectural Economic Accounts	Annual	Higher is better	T1
Total population Total resident population of the municipality.	people	e-Stat — System of Social and Demographic Statistics (municipal data)	Annual	Higher is better	T1
Annual income (individual) Average annual pay (salary + bonus) of one full-time worker. Not household income.	¥	MHLW — Basic Survey on Wage Structure	Annual	Higher is better	T1
Household assets (avg.) Average savings balance per household, before subtracting debt.	¥	MIC — National Survey of Family Income, Consumption and Wealth	Every 5 years	Higher is better	T1
Individual income (median) The midpoint of workers’ annual income — not skewed by high earners.	¥	MIC — Employment Status Survey	Every 5 years	Higher is better	T1
Household savings (median) The midpoint of household savings — a representative figure not skewed by the wealthy.	¥	MIC — National Survey of Family Income, Consumption and Wealth	Every 5 years	Higher is better	T1
Fiscal capability index (municipality) How much of its required spending a municipality can fund from its own tax revenue. Above 1.0 means it relies on no national local-allocation tax — a financially independent government.	index	MIC — Key Fiscal Indicators of Municipalities (Local Public Finance Survey)	Annual	Higher is better	T1
Current-account ratio Share of recurring revenue consumed by fixed recurring costs. Lower means more room in the budget for discretionary spending.	%	MIC — Key Fiscal Indicators of Municipalities (Local Public Finance Survey)	Annual	Lower is better	T1
Real debt-service ratio Share of the budget taken up by annual debt repayment. Higher means the budget is more constrained by debt.	%	MIC — Key Fiscal Indicators of Municipalities (Local Public Finance Survey)	Annual	Lower is better	T1
Jobs-to-applicants ratio How many openings exist per job seeker at Hello Work offices. 1.0 means openings equal seekers; higher means jobs are easier to find.	×	Statistics Bureau, MIC — Social Indicators by Prefecture (F: Labour)	Annual	Higher is better	T1
Dual-income households Share of households where both spouses work. A contextual metric shaped by geography and industry, with no better-or-worse.	%	Statistics Bureau, MIC — Social Indicators by Prefecture (F: Labour)	Annual	Context-dependent	T1
Outstanding local-bond ratio How many times the standard annual budget a prefecture’s outstanding borrowing equals. Higher means debt repayment weighs more heavily on future budgets.	%	Statistics Bureau, MIC — Social Indicators by Prefecture (D: Administrative Base)	Annual	Lower is better	T1
Fiscal capability index How much of its required spending a prefecture can fund from its own tax revenue. Above 1.0 means it relies on no national local-allocation tax — a financially independent government.	index	Statistics Bureau, MIC — Social Indicators by Prefecture (D: Administrative Base)	Annual	Higher is better	T1
Household income (median) The midpoint of annual household income — total household income, distinct from individual income.	¥	MIC — National Survey of Family Income, Consumption and Wealth	Every 5 years	Higher is better	T1
Prefectural assembly pay (monthly) Ordinance-set monthly pay of a prefectural assembly member (rank-and-file). A baseline for the check-and-balance role.	¥/mo	National Association of Chairpersons of Prefectural Assemblies — Survey on Assembly-Member Compensation + each prefectural assembly’s compensation ordinance	Annual	Lower is better	T1
Designated-city assembly pay (monthly) Ordinance-set monthly pay of a rank-and-file member of a designated-city assembly (the 20 ordinance-designated cities).	¥/mo	National Association of Chairpersons of City Councils — Survey on City-Council Member Compensation (as of 31 Dec 2024) + each designated-city council ordinance	Annual	Lower is better	T1
Assembly pay ÷ residents The assembly’s total annual member pay divided per resident. A way to view the scale of the assembly from a resident’s perspective.	¥/person/yr	National Association of Chairpersons of Prefectural Assemblies survey + assembly seat counts + e-Stat Population Census (population)	Annual	Lower is better	T1
Assembly pay ÷ resident income How many times a member’s annual pay equals the average resident’s annual income. A multiple that compares pay against local income levels.	×	Prefectural assembly-member annual income (monthly×12 + 4 months’ year-end allowance) ÷ e-Stat taxable income per capita	Annual	Lower is better	T1
Establishments Total number of business establishments in the municipality (2021 Economic Census). A measure of the scale of industrial concentration, not the size of any single establishment.	事業所	e-Stat 経済センサス-活動調査（事業所・従業者）	不定期（経済センサス・概ね5年）	Higher is better	T1
Workers Total number of people working at all establishments in the municipality (2021 Economic Census). A gauge of the scale of local employment.	people	e-Stat 経済センサス-活動調査（事業所・従業者）	不定期（経済センサス・概ね5年）	Higher is better	T1

Childcare(15)

Metric	Unit	Source	Update	Reading	Tier
Total fertility rate Estimated number of children a woman bears in her lifetime. Higher means more.	—	MHLW — Vital Statistics	Annual	Higher is better	T1
Daycare waitlist rate Share of children who applied for daycare but could not get a place. Lower means easier to enroll.	%	Children and Families Agency — Status of Childcare Facilities	Annual	Lower is better	T1
Daycare enrollment rate Share of children aged 0–5 enrolled in daycare (the actual-use side).	%	Statistics Bureau, MIC — Social Indicators by Prefecture (E: Education)	Annual

Education(4)

Metric	Unit	Source	Update	Reading	Tier
Public elementary spend per pupil The actual education spending a local government invests per public-elementary pupil.	¥	Statistics Bureau, MIC — Social Indicators by Prefecture (D: Administrative Base)	Annual	Higher is better	T1
Elementary long-absence rate Share of elementary pupils absent 30+ days a year due to non-attendance (per 1,000).	‰	Statistics Bureau, MIC — Social Indicators by Prefecture (E: Education)	Annual	Lower is better	T1
University-graduate share Share of residents whose highest degree is university or graduate school. A contextual metric shaped by industry, with no better-or-worse.	%	Statistics Bureau, MIC — Social Indicators by Prefecture (E: Education)

Healthcare(11)

Metric	Unit	Source	Update	Reading	Tier
Healthy life expectancy (women) Years spent without limitation in daily life. Longer means healthier aging.	years	MHLW Health Sciences Research — Healthy Life Expectancy (Health Japan 21)	Every 3 years	Higher is better	T1
Life expectancy Estimated years a newborn will live. Higher means greater longevity.	years old	MHLW — Life Tables by Prefecture	Every 5 years	Higher is better	T1
Doctors (per 100k) Doctors per 100,000 people. More means easier access to medical care.	people	Statistics Bureau, MIC — Social Indicators by Prefecture

Housing(12)

Metric	Unit	Source	Update	Reading	Tier
Home-ownership rate Share of households that own their home. Higher means more owner-occupied housing.	%	MIC — Housing & Land Survey	Every 5 years	Higher is better	T1
House price / m² (median) Median transaction price per m² for detached houses. Based on land + building total.	¥10k/m²	MLIT — Real Estate Information Library (REINFOLIB)	Quarterly	Lower is better	T1
Used condo price / m² (median) Median transaction price per m² for used condominiums.	¥10k/m²	MLIT — Real Estate Information Library (REINFOLIB)	Quarterly

Safety(4)

Metric	Unit	Source	Update	Reading	Tier
Recorded crimes (per 1,000) Recorded crimes per 1,000 people. Fewer means safer.	cases	NPA — Crime Statistics	Annual	Lower is better	T1
Traffic accidents (per 100k) Traffic accidents per 100,000 people. A road-safety metric independent of criminal offenses.	cases	Statistics Bureau, MIC — Social Indicators by Prefecture (K: Safety)	Annual	Lower is better	T1
Fires (per 100k) Fire outbreaks per 100,000 people. A proxy for housing and building safety.	cases	Statistics Bureau, MIC — Social Indicators by Prefecture (K: Safety)

Culture & Leisure(2)

Metric	Unit	Source	Update	Reading	Tier
Museums (per million) Museums per million people. A density of cultural resources, independent of libraries.	facilities	Statistics Bureau, MIC — Social Indicators by Prefecture (G: Culture & Sports)	Annual	Higher is better	T1
Public gyms (per million) Public gymnasiums per million people. The supply of indoor exercise facilities.	facilities	Statistics Bureau, MIC — Social Indicators by Prefecture (G: Culture & Sports)	Annual	Higher is better	T1

Climate(12)

Metric	Unit	Source	Update	Reading	Tier
Forest-area share Share of prefectural land covered by forest. The extent of the wider natural environment.	%	Statistics Bureau, MIC — Social Indicators by Prefecture (B: Natural Environment)	Annual	Higher is better	T1
Annual sunshine hours Annual cumulative sunshine hours observed by JMA. A measure of outdoor and daylight conditions.	hours	Statistics Bureau, MIC — Social Indicators by Prefecture (B: Natural Environment)	Annual	Higher is better	T1
Annual average temperature Annual average temperature observed by JMA. Preference for warm or cool is personal, so there is no better-or-worse.	°C	Statistics Bureau, MIC — Social Indicators by Prefecture (B: Natural Environment)

Population & Vitality(3)

Metric	Unit	Source	Update	Reading	Tier
Population change rate Year-on-year rate of population change. Positive is growth, negative is decline.	%	MIC — Population Estimates	Monthly	Higher is better	T1
Under-15 population share Share of the total population under 15. A gauge of the presence of child-rearing generations.	%	Statistics Bureau, MIC — Social Indicators by Prefecture (A: Population & Households)	Annual	Higher is better	T1
Net migration rate Net of in-migrants minus out-migrants per 1,000 people. Net inflow or outflow is neither better nor worse.	‰	Statistics Bureau, MIC — Social Indicators by Prefecture (A: Population & Households)

How the 6-axis score and standardized score in "Search by conditions" are computed

In the top-page "Search by conditions", candidates are first screened out by hard conditions (budget, commute ceiling, exclusion of fatal hazards), and the remaining neighborhood blocks (chōchō) are then ranked by a 6-axis score weighted to how much you care about each. Screening and ranking are separate processes. Each axis score is computed by the fixed formula below; no AI inference is used. The standardized score is the relative position within the candidate set (mean 50), not an absolute rating.

Axis	What it measures / how it is computed	Data source	Coverage
Commute	Travel time to the destination station. Computed along the shortest route traced through the rail network, scored on an absolute basis of 0 min = full marks down to 120 min = 0. Points with no route are filled with straight-line distance ×2.	MLIT National Land Numerical Information — Railways	In place (verify exact times with a transit app)
Affordability	Converts the unit price of second-hand properties at that point into a cheapness ranking within its regional bloc (Kanto, Kansai, etc.). For condominiums, total transaction price ÷ exclusive floor area; for detached houses, total transaction price (land + building) ÷ land area — both a unit price dividing the building-inclusive total by area. Cheaper scores higher, and the value does not move as you change search conditions.	Reinfolib real-estate transaction prices (actual closing prices)	In place (points with actual transaction prices; thin points use the municipal average of the same type as a reference)
Disaster safety	Pre-computed intersection with hazard areas for flooding, storm surge, tsunami, and the like. The more a point intersects, the more it is docked.	MLIT hazard maps (overlaid)	Under development (only part of flooding reflected; landslide / liquefaction not yet covered). Always check the official hazard map.
Sidewalks	A walkability score from sidewalk coverage and the density of nearby facilities (POIs).	OpenStreetMap / MLIT National Land Numerical Information	Under development (un-ingested points are reference values)
Liquidity	Transaction activity (a guide based on the time from listing to closing). More active scores higher.	Reinfolib transaction data	Under development (un-ingested points are reference values)
Surroundings	A childcare environment computed from the childcare-waitlist rate, share of child-rearing households, and the like (inheriting municipal-level delivered values).	e-Stat (childcare waitlists / population census)	In place (municipal grain)

The score breakdown for each neighborhood block can be checked one axis at a time, together with the basis (raw data), by opening the ranking card. Axes with un-ingested data are clearly marked "reference value" and are not shown as if real data existed. The score is a starting point, not a conclusion. For individual property decisions, always combine it with on-the-ground and professional checks.

Why we do not add "proximity to a development plan" to the score

It is natural to think "if there is a large development plan nearby, asset value should be higher". But checking the peer-reviewed literature on real-estate price formation and urban planning against primary sources (the Japan Association for Real Estate Sciences, Kobe University, MLIT, overseas peer-reviewed journals, and others), we found that scoring the effect of development as added points instead misleads the judgement. As a result of that verification, this site does not add development plans to the score; it shows them only as context, on the map heatmap and in the "basis for asset value" panel. There are three reasons.

It double-counts — most of the price rise happens at the time a plan is announced(one study found about 97% of the total rise is priced in at announcement). By the time a development reaches our data, the effect is already reflected in transaction prices. Because this site evaluates the observed transaction prices themselves on the "affordability" axis, separately scoring proximity to development would count the same effect twice.
The sign reverses between city and country— the effect of a Shinkansen or new line pushes prices up in large cities where trading is thick, while in thin markets there are examples of decline (one study found +7.8% in the north and −9.2% to −14.3% in the south on the same line). Uniformly adding points for "near = plus" can steer thin markets in the wrong direction.
The effect depends on quality and distance, and does not arise automatically — the impact on prices is local, roughly disappearing within about 1.5 km of a station, and some lines show no significant effect. Proximity alone cannot be uniformly turned into a score.

How future development will play out cannot be measured from past cases alone (unprecedented mega projects, for instance). We do not quantify that part; we show the facts — a plan's location, type, status, and source — on the map and in the basis panel, and leave the judgement to the reader. Sources: MLIT National Land Numerical Information, Location Optimization Plan areas (A50) / operator and municipal disclosures / peer-reviewed literature on real-estate price formation (J-STAGE Japan Association for Real Estate Sciences, Kobe University, Board of Audit, MLIT, Springer, ScienceDirect, and others).

Frequently asked questions

What does it mean to separate "potential" from "delivered outcome"?›

The Fiscal Capacity Index (how far tax revenue can cover administrative costs) shows the depth of a municipality's pockets. Childcare waitlists and daycare capacity show how far a service is actually reaching people. They wear similar faces, but they answer different questions. Force them into a single "overall score" and the counter-intuitive facts disappear — "plenty of money yet a long childcare waitlist", or "thin budget yet generous support". So this site lays them on separate axes and shows the tensions as they are.

Why only the 47 prefectures, rather than international comparison?›

The 47 prefectures are the largest grain at which the population census, economic accounts, and fiscal statistics line up under the same definitions. International comparison introduces gaps in currency, prices, and statistical definitions, which makes a same-basis side-by-side effectively hard. First, line everything up at a consistent grain — Phase 1 concentrates there, Phase 1.5 extends to municipalities (about 1,920), and Phase 2 onward adds international reference values.

Who chooses the metrics, and on what basis?›

CPA Atlas examines and adopts each one individually. The single test is whether the metric truly represents the concept it claims to. No single metric can sum up "ease of raising children", so childcare waitlists, daycare capacity, child allowances, and average birth rate are laid out independently. The log of what was adopted and rejected is planned for publication under the docs directory on GitHub.

How fresh is the data?›

It follows the release cycle of each public statistic (population census every 5 years, economic census every 5 years, labour-force survey quarterly, and so on). Each metric's update frequency is stated in the catalogue on this page, and the date of ingestion is monitored separately. "Living" metrics such as weather and earthquakes update automatically from official APIs at intervals of 5 minutes to 1 hour.

Is there anything that worries you about its use for investment decisions?›

Yes. What is laid out here is a secondary visualisation of public statistics; it is not usable for buying or selling individual properties or valuing individual companies. What I saw repeatedly on audit engagements is that deals decided on numbers alone, with the feel of the ground stripped out, collapse badly later. Statistics are a starting point, not a conclusion. For sale timing and property selection, always draw on the site and on professionals on the ground.