Japan Off-Peak

Blog › 2026-10-30

Is Hiroshima crowded during Golden Week? May exceeds average

By Japan Off-Peak (TechAthletes) · 2026-10-30 · Data: Japan Tourism Agency, counted by us (method)

Key points

Is Hiroshima crowded during Golden Week? That exact question was the second of ten Google autocomplete suggestions for “is hiroshima crowded” in our English, US check on 15 October 2026. To examine the available evidence, we counted hotel and other lodging stays from official statistics. We did not count people at attractions, on trains or in streets. The geographical coverage is Hiroshima Prefecture, including Miyajima/Hatsukaichi, rather than Hiroshima city alone.

Short answer: May was above the average month for lodging stays in Hiroshima Prefecture in each year examined. In 2025, May recorded 34,664 guests per day, ranked fourth among the year’s months, and reached 108.9% of the average month. That supports describing May as a comparatively busy lodging month. It does not establish a Golden Week crowd level: the holiday crosses the April–May boundary, while these figures cover all of May. (Source: Japan Tourism Agency, confirmed annual results)

How May compares across three years

Guests per day means a month’s guest nights divided by the number of days in that month.

This measure lets us compare months of different lengths without treating a longer month as busier simply because it contains more days. It describes overnight stays averaged across the month, rather than a count of distinct travelers.

The average month is the simple average of a year’s twelve monthly guests-per-day figures.

The table below reproduces selected columns from our calculations of the agency’s confirmed annual results. Each rank compares May with the other months in the same year.

Year All guests per day Rank (of 12) vs the average month
2023 36,219 3 114.4%
2024 35,161 5 106.6%
2025 34,664 4 108.9%

Source: Japan Tourism Agency, Overnight Travel Statistics Survey, confirmed annual results for 2023, 2024 and 2025.

May was above its own year’s average in every case, but it was never the busiest month. The highest daily lodging total occurred in August in 2023 and November in both 2024 and 2025. January had the lowest total in each year.

The distinction between the absolute count and the annual ranking matters. May’s daily total was 35,161 in 2024 and 34,664 in 2025, yet its rank moved from fifth to fourth. A rank describes a month’s position within that particular year; it does not by itself show that the number of stays increased.

The national comparison also gives a different picture from the prefectural one. Across Japan, May reached 98.0% of the average month in 2023, 97.4% in 2024 and 99.9% in 2025. Hiroshima Prefecture’s corresponding figures were 114.4%, 106.6% and 108.9%. National seasonality therefore does not substitute for the prefecture’s own results.

For the wider seasonal context, see Hiroshima month by month. The May comparison across prefectures provides another way to examine the month without assuming every destination follows the same pattern.

Foreign and domestic stays follow different peaks

Foreign means lodging guests with no address in Japan; foreign residents of Japan count as domestic guests.

Domestic stays are calculated as all stays minus foreign stays. These categories therefore describe residence, rather than simply dividing guests by nationality.

In May 2023, foreign guests accounted for 3,931 stays per day, ranking seventh and reaching 99.6% of their annual monthly average. Domestic guests accounted for 32,289 per day, ranking second and reaching 116.5%. The domestic component was above its average while the foreign component was close to its average.

In 2024, both categories ranked fifth in May. Foreign stays reached 117.7% of their average month, while domestic stays reached 104.5%. In 2025, both ranked fourth, at 126.4% and 105.0% respectively. Matching ranks do not mean the categories stood equally far above their own averages.

Their annual peaks also differed. Domestic stays peaked in August in all three years. Foreign stays peaked in October in 2023 and 2024, then in April in 2025. May was therefore not the foreign or domestic peak in any of these years.

The nationality breakdown adds detail, but covers only lodgings with ten or more staff. Its shares use that narrower table’s total, including nationality unknown, and must not be treated as shares of all foreign stays across every lodging.

The United States was the largest individually named country in the supplied May breakdown in each year, with shares of 17.8% in 2023, 13.9% in 2024 and 12.5% in 2025. In 2025, the other leading named countries were the United Kingdom at 8.9%, Australia at 8.7%, France at 7.9% and Germany at 7.4%. “Other countries” accounted for 20.1%, but that is a combined category rather than a single nationality.

Source: Japan Tourism Agency, confirmed annual results and reference table 1.

Room occupancy varies by lodging type

Room occupancy is the proportion of lodging rooms occupied, expressed as a percentage.

It measures a different aspect of accommodation use from guests per day. The following figures are room occupancy rates from the agency’s table 8; every lodging-type value is a percentage.

Year All lodgings Ryokan Resort hotels Business hotels City hotels
2023 65.5 45.7 74.1 71.3 65.7
2024 64.4 45.6 66.6 71.8 75.0
2025 67.5 36.3 63.8 79.7 74.8

Source: Japan Tourism Agency, confirmed annual results, table 8.

The overall figure does not describe every type of accommodation equally. In May 2025, business hotels recorded 79.7% occupancy and city hotels 74.8%, while resort hotels recorded 63.8% and ryokan 36.3%. These are category averages across the prefecture, not availability checks for individual properties.

May’s all-lodging occupancy ranked fifth in 2023, sixth in 2024 and fourth in 2025. The highest occupancy month was November in 2023 and 2024, and October in 2025. That differs from the guests-per-day peak in some years, reinforcing the need to keep the two measures separate.

For someone choosing accommodation, the table provides historical context for the type of lodging being considered. It cannot establish whether a particular hotel has a room available for a particular holiday night, or what that room will cost.

Hiroshima’s May weather normals

The Japan Meteorological Agency’s 1991–2020 normals for the Hiroshima observation station provide a weather reference alongside the lodging figures. These are long-term normals, not a forecast for a Golden Week trip.

May’s mean temperature is 19.6 °C, with an average daily high of 24.4 °C and an average daily low of 15.1 °C. Monthly precipitation is 169.8 mm, and sunshine totals 210.8 hours.

For comparison, April’s mean temperature is 14.8 °C and its precipitation is 141.0 mm. June’s mean temperature is 23.2 °C and its precipitation is 226.5 mm. May sits between those neighboring months on both measures.

These observations add weather context to the choice of month. They do not explain the lodging patterns above: the supplied evidence does not establish a causal relationship between weather and the number of stays. Nor does a monthly precipitation normal specify which dates will be wet.

Source: Japan Meteorological Agency, Hiroshima, 1991–2020 normals, retrieved 2026-10-16.

What this data can't tell you

The central limitation is the calendar. Golden Week runs from late April into early May, while the hotel figures examined here cover May as a whole. Calling them “Golden Week hotel figures” would assign them a period they do not measure.

Monthly averages cannot identify the busiest holiday date, distinguish weekends from weekdays, or show whether the first half of May was busier than the second. They also cannot show whether a lower whole-month total coincided with a concentrated peak on particular days.

Geography imposes another limit. These are Hiroshima Prefecture results, including Miyajima/Hatsukaichi. They do not describe the conditions on a particular Hiroshima city street, at a temple, or at an individual attraction. A prefectural lodging total cannot establish a queue length or the experience of walking through a specific place.

The measurement itself has limits too. Lodging stays are a proxy for crowd pressure, not a direct count of people in public places. Nationality shares from larger establishments cannot be extended to all accommodation, and occupancy averages cannot be read as a booking inventory.

The supported conclusion is consequently specific: May was above the average month for daily lodging stays across the prefecture in the years examined. The evidence does not support a precise answer about Golden Week itself, nor a claim about why the monthly totals changed. Our explanation of how we count sets out the method behind these comparisons.

Sources

How this article was made: an AI model drafted the text from our own tabulation of the official tables. Before publishing, every number in it was checked automatically against that tabulation, and the draft was reviewed by a second AI model for unsupported claims.

Related on Japan Off-Peak

Japan Off-Peak newsletter. Get one new count from Japan's official data each week, free.

Form not showing? Subscribe on Substack.