The Atarashift Blog
Insights for working in Japan
Guides, interviews, and honest advice for international talent building a career in Japan.
The Atarashift Blog
Guides, interviews, and honest advice for international talent building a career in Japan.
Kyoto records 6.23 foreign guest-nights per resident, five times Japan's average. The full 47-prefecture ranking, plus which countries keep coming back.
Kyoto is Japan's most tourist-dense prefecture. In 2025 it recorded 6.23 foreign guest-nights for every resident, against a national average of 1.25. Tokyo, which takes more foreign guest-nights than anywhere else in absolute terms, ranks third at 3.59.
We worked out the per-resident figures below ourselves. We joined the Japan Tourism Agency's guest-night table to the 2025 census head count. The Tourism Agency publishes guest-nights, and the census publishes population, but neither publishes the ratio. As far as we can tell, the full 47-prefecture ranking below is not available anywhere else.
That gap between "most visitors" and "most visitors per person" is the whole point. Raw visitor counts tell you where Japan's tourism economy sits. Per-resident counts tell you where it is actually felt, and the two rankings disagree sharply. This post gives the full 47-prefecture list and explains which places move most between the two measures. It also covers the other half of the picture: which countries treat Japan as a repeat destination rather than a once-in-a-lifetime trip.
Kyoto, at 6.23 foreign guest-nights per resident in 2025. Okinawa is second at 4.24, Tokyo third at 3.59, Osaka fourth at 2.64 and Hokkaido fifth at 2.25.
A foreign guest-night is one night that one foreign visitor spent in paid accommodation in that prefecture. It is not a headcount of visitors. One person staying four nights counts as four guest-nights. Japan recorded 154.3 million foreign guest-nights in 2025 across a resident population of 123.0 million.
Dividing one by the other gives a rough measure of tourism pressure: how much visitor activity each prefecture absorbs per resident.
The full ranking for 2025. "Rank by volume" is where each prefecture sits on the raw guest-night count, for comparison.
| Rank | Prefecture | Per resident | Guest-nights | Population | Rank by volume |
|---|---|---|---|---|---|
| 1 | Kyoto | 6.23 | 15,591,430 | 2,502,747 | 3 |
| 2 | Okinawa | 4.24 | 6,230,110 | 1,468,220 | 6 |
| 3 | Tokyo | 3.59 | 51,138,750 | 14,246,219 | 1 |
| 4 | Osaka | 2.64 | 23,179,170 | 8,764,578 | 2 |
| 5 | Hokkaido | 2.25 | 11,210,990 | 4,985,419 | 4 |
| 6 | Yamanashi | 2.16 | 1,686,860 | 779,912 | 15 |
| 7 | Ishikawa | 1.69 | 1,834,510 | 1,088,221 | 14 |
| 8 | Fukuoka | 1.54 | 7,831,810 | 5,081,879 | 5 |
| 9 | Oita | 1.02 | 1,097,070 | 1,076,875 | 18 |
| 10 | Kagawa | 1.02 | 923,840 | 907,725 | 20 |
| 11 | Nagano | 0.94 | 1,841,490 | 1,954,950 | 13 |
| 12 | Kumamoto | 0.93 | 1,565,840 | 1,678,090 | 17 |
| 13 | Gifu | 0.84 | 1,591,240 | 1,891,489 | 16 |
| 14 | Hiroshima | 0.71 | 1,899,630 | 2,683,399 | 11 |
| 15 | Chiba | 0.69 | 4,339,310 | 6,258,512 | 8 |
| 16 | Wakayama | 0.67 | 583,130 | 864,262 | 22 |
| 17 | Aichi | 0.61 | 4,531,530 | 7,449,403 | 7 |
| 18 | Shizuoka | 0.55 | 1,914,560 | 3,468,845 | 10 |
| 19 | Kanagawa | 0.45 | 4,154,890 | 9,193,657 | 9 |
| 20 | Aomori | 0.45 | 508,080 | 1,140,395 | 26 |
| 21 | Miyagi | 0.44 | 971,080 | 2,227,240 | 19 |
| 22 | Nagasaki | 0.43 | 529,940 | 1,232,190 | 24 |
| 23 | Iwate | 0.38 | 429,950 | 1,125,502 | 29 |
| 24 | Ehime | 0.38 | 475,450 | 1,260,088 | 27 |
| 25 | Kagoshima | 0.37 | 562,520 | 1,512,969 | 23 |
| 26 | Toyama | 0.37 | 361,290 | 985,675 | 31 |
| 27 | Hyogo | 0.36 | 1,897,870 | 5,323,825 | 12 |
| 28 | Nara | 0.35 | 442,530 | 1,269,180 | 28 |
| 29 | Okayama | 0.35 | 625,790 | 1,808,664 | 21 |
| 30 | Saga | 0.31 | 240,330 | 781,214 | 39 |
| 31 | Tokushima | 0.30 | 200,850 | 675,489 | 40 |
| 32 | Yamagata | 0.29 | 285,370 | 993,127 | 35 |
| 33 | Tottori | 0.29 | 149,680 | 523,732 | 43 |
| 34 | Niigata | 0.25 | 519,340 | 2,068,476 | 25 |
| 35 | Tochigi | 0.21 | 384,230 | 1,864,833 | 30 |
| 36 | Shiga | 0.20 | 273,820 | 1,392,439 | 36 |
| 37 | Fukushima | 0.19 | 325,210 | 1,711,937 | 32 |
| 38 | Mie | 0.19 | 317,050 | 1,694,896 | 33 |
| 39 | Miyazaki | 0.19 | 188,790 | 1,018,904 | 41 |
| 40 | Kochi | 0.18 | 115,950 | 643,437 | 45 |
| 41 | Gunma | 0.16 | 303,710 | 1,867,582 | 34 |
| 42 | Shimane | 0.14 | 90,140 | 629,460 | 47 |
| 43 | Akita | 0.14 | 125,710 | 882,100 | 44 |
| 44 | Fukui | 0.13 | 97,840 | 729,386 | 46 |
| 45 | Yamaguchi | 0.12 | 156,380 | 1,264,006 | 42 |
| 46 | Ibaraki | 0.10 | 270,850 | 2,791,207 | 37 |
| 47 | Saitama | 0.04 | 258,210 | 7,287,169 | 38 |
National figure: 1.25 guest-nights per resident.
Twelve prefectures move nine places or more between the two measures. The pattern is not random, and it comes down to how many people live there.
Prefectures with one major attraction and few residents climb. Yamanashi goes from 15th by volume to 6th per resident. It has 779,912 people and most of Mount Fuji's northern approach, including the Fuji Five Lakes. Oita climbs from 18th to 9th on the strength of Beppu and Yufuin. Those two onsen towns sit in a prefecture of just over a million people. Kagawa and Tottori each climb 10 places.
Large commuter prefectures fall, for the same reason in reverse. Hyogo drops 15 places, from 12th to 27th, the biggest fall on the list. Kobe draws real visitor numbers, but Hyogo also holds 5.3 million residents, many of whom commute into Osaka. Kanagawa and Aichi each fall 10 places for the same reason. Saitama, with 7.3 million residents and 258,210 foreign guest-nights, finishes last.
The very top barely moves. Kyoto, Okinawa, Tokyo, Osaka and Hokkaido occupy the top five on both measures in some order. Between them they account for 69.6% of all foreign guest-nights in Japan. Tourism in Japan is concentrated, and dividing by population does not change that.
The second half of the picture is repeat visits. Japan received a record 42.68 million foreign visitors in 2025, but "a visitor" means something very different depending on where they flew in from.
In the Japan Tourism Agency's consumption survey, 38.8% of all surveyed visitors had already been to Japan at least once in the previous 12 months. Broken down by market, the spread is enormous.
| Market | Visited within past 12 months | Average nights | Spend per trip |
|---|---|---|---|
| Hong Kong | 65.2% | 6.2 | ¥213,751 |
| Taiwan | 55.7% | 6.5 | ¥160,079 |
| South Korea | 50.2% | 4.0 | ¥98,963 |
| Singapore | 41.9% | 9.1 | ¥312,500 |
| Thailand | 39.1% | 8.6 | ¥181,237 |
| Philippines | 37.9% | 15.5 | ¥175,998 |
| China | 35.0% | 9.1 | ¥229,011 |
| Indonesia | 25.1% | 15.3 | ¥191,415 |
| Malaysia | 24.5% | 8.8 | ¥201,092 |
| U.K. | 18.8% | 14.5 | ¥348,726 |
| Germany | 18.1% | 18.0 | ¥372,025 |
| Australia | 17.7% | 13.5 | ¥343,693 |
| U.S. | 17.6% | 12.0 | ¥315,101 |
| Canada | 17.4% | 14.6 | ¥306,305 |
| France | 14.1% | 18.4 | ¥339,372 |
| Italy | 11.4% | 13.7 | ¥290,577 |
| Spain | 8.2% | 13.9 | ¥307,076 |
So there is a clear break in the middle of that table. Every market inside roughly four hours' flying time sits above 24%. Every long-haul market sits below 19%. For Hong Kong, Taiwan and South Korea, a trip to Japan resembles how a German might treat a weekend in Austria.
It is tempting to read that spend column and conclude some nationalities are simply higher-value visitors. But that reading does not survive contact with the nights column.
South Korean visitors spend ¥98,963 per trip. Australians spend ¥343,693, about 3.5 times more. But Koreans stay 4.0 nights on average and Australians stay 13.5. Divide spend by nights and the two are almost identical: ¥24,991 a night for Korea, ¥25,497 for Australia.
Trip frequency does not rescue the "high-value visitor" framing either. A departure-gate survey samples a frequent traveller once per trip. Correcting for that, Koreans average roughly 1.46 trips a year against 1.11 for Australians. That still leaves annual spend well apart. What really differs is total time in the country: about 5.8 nights a year for a Korean visitor against 14.9 for an Australian.
So per-trip spending is mostly a measure of how long someone can stay. Short trips are cheap trips, and short trips are what you take when the flight is two hours and you can come back next quarter.
Tourism density is not just a number if you are choosing where to live. A few things it changes.
Rent and housing pressure. Places at the top of this list compete with short-term rentals for the same buildings. Kyoto in particular has spent years tightening rules on minpaku (private lodging) precisely because of this. If you are weighing up cities, check the local short-term rental rules as well as rent. Our cost of living breakdown for Tokyo covers what the numbers look like in the capital.
Commuting and daily friction. Kyoto's buses and Kyoto Station get crowded in a way that a raw visitor count does not show. A city of 2.5 million absorbing 15.6 million foreign guest-nights feels different from Tokyo absorbing 51.1 million across 14.2 million residents.
Where hospitality and service jobs are. The flip side of density is jobs. Hotels and restaurants are among the sectors that rely most on foreign workers in Japan. Demand sits in exactly the places at the top of this list. If your Japanese is still developing, tourism-facing regions often have more roles where English and a third language are genuinely useful. See our breakdown of which jobs are most in demand in Japan.
The quiet 40. Thirty-eight prefectures sit below the national average of 1.25. If you want life in Japan without the crowds, the bottom half of this table is a reasonable starting shortlist. Weigh it against job availability and transport links.
For the same data ranked by raw volume and split by nationality, see where foreign visitors actually go in Japan.
One guest-night is one person staying one night in paid accommodation. A visitor who stays five nights in Kyoto counts as five Kyoto guest-nights. It counts nights stayed rather than people, which makes it a better gauge of tourism pressure than an arrival count.
Japan does not publish visitor headcounts at prefecture level. JNTO arrival figures are national, and split by port of entry rather than by where people went. Guest-nights from the Accommodation Survey are the only figures we have by prefecture.
By this measure, yes, and by a factor of about 1.7. Kyoto absorbs 6.23 foreign guest-nights per resident against Tokyo's 3.59. Tokyo takes more visitors in total, but spreads them across far more residents and a much bigger city.
No. These figures cover foreign guest-nights only. Domestic travel is large in Japan. It would change the picture in places like Nagano and Shizuoka, which draw heavily from inside the country.
By repeat rate, Hong Kong. In 2025, 65.2% of surveyed Hong Kong visitors had already been to Japan within the previous 12 months. By total arrivals, South Korea led with 9.46 million visitors, followed by China with 9.10 million.
Guest-night figures come from the Japan Tourism Agency's Accommodation Survey (宿泊旅行統計調査), 2025 annual final figures, retrieved 31 August 2026. The prefecture-by-nationality table covers hotels with 10 or more employees. That is a smaller group than the headline all-hotel count, so small ryokan and guest houses are missed. That matters most in rural areas, where small firms make up more of the market. It likely makes the bottom of the table look a little low.
Population figures are from the 2025 Population Census (令和7年国勢調査) preliminary count, as of 1 October 2025, retrieved 31 August 2026. The 47 prefecture figures sum exactly to the published national total of 123,049,524.
Repeat-visit, spend and nights figures come from the Japan Tourism Agency's Inbound Consumption Trend Survey (インバウンド消費動向調査), retrieved 31 August 2026. These are 2025 annual final (確報) figures, tables 1-1, 3-1 and 4-1. Note that preliminary (速報) figures released in January 2026 and widely quoted in the press differ from these final numbers. The survey covers general visitors on all trip purposes and excludes cruise passengers.
Arrival totals are from JNTO's visitor arrivals statistics, retrieved 31 August 2026.
The per-resident column is our own calculation from those two datasets and is not published by either agency.
Allowances are 15.9% of the average Japanese pay packet, and most of them do not count toward your bonus or your overtime rate. How to read a Japanese offer.
Housing is where foreign residents report the most discrimination in Japan. What you pay upfront, how guarantor companies screen you, and how to get approved.
Japan caps overtime at 45 hours a month, backed by criminal penalties. Full-time workers average 13.2 hours, but transport runs three times manufacturing.
