How the quietness estimate works, and what it cannot hear
The quietness score on SHIZUAN did not come from a sound meter. It combines public data about the things that usually make a place loud into a single number from 0 to 100. This guide explains how that number is built and, just as importantly, what it cannot pick up.
Why an estimate?
Japanese authorities do monitor road traffic noise, and the results are published, mostly as maps and images. What we could not find, as of September 2026, is noise data that can be compared numerically across every neighborhood in Tokyo. So instead of measuring sound, SHIZUAN measures how much of each neighborhood is taken up by typical sources of sound.
Four ingredients and their weights
The score starts at 100 and loses points for four ingredients. The more of them a neighborhood has, the lower (livelier) its score.
- Zoning (35%). The share of the neighborhood's land zoned commercial, neighborhood commercial or industrial, from the Tokyo Metropolitan Government's city planning data. Mostly residential zoning keeps the score high.
- Late-night venues (30%). Bars, pubs, nightclubs, karaoke boxes and similar places listed in OpenStreetMap, per square kilometer. Izakaya, the Japanese pubs that double as restaurants, count as half.
- Major transport (25%). The share of the neighborhood inside a 60 m corridor along expressways, national routes and other major roads, a 40 m corridor along railways that run above ground, or within 3 km of an airport. Underground railway sections are left out; road tunnels are still counted.
- Station ridership (10%). The total daily ridership of stations within 500 m of the neighborhood's center. More people coming and going lowers the score.
Venue density and ridership vary enormously, so both are put on a logarithmic scale and compared with fixed reference values: 600 venues per square kilometer and 1,400,000 riders per day. Because those references never change, the score means the same thing in every region the map may cover in the future.
Five fixed bands
Scores are grouped into five levels with fixed boundaries:
| Level | Estimated quietness score |
|---|---|
| Very lively | 0 – 40 |
| Lively | 40 – 55 |
| Mixed | 55 – 75 |
| Quiet | 75 – 88 |
| Very quiet | 88 – 100 |
When you choose the quietness tab on the map, column height shows liveliness (100 minus the score), and ripples spread from the liveliest neighborhoods. Deeper teal means quieter.
What the estimate cannot hear
- The inside of a building. Window glazing, the direction a room faces and the floor it is on change what you actually hear far more than the neighborhood does.
- Temporary noise. Construction, events, seasonal festivals and emergency sirens are not in the data.
- Unlisted venues. Only bars and similar places recorded in OpenStreetMap are counted.
- Flight paths. Airports are handled as a 3 km circle. Areas under approach and departure routes further away are not captured.
- The quiet street behind the main road. Because the score works on shares of each neighborhood's area, a calm side street and a busy avenue in the same block get the same number.
About the weights and reference values
The weights (35, 30, 25 and 10) are a design choice made by this site. They have not been calibrated against noise measurements, which is one more reason to treat the score as a rough guide. The two reference values are not arbitrary: they are the 99th percentile of the non-zero values found in Tokyo, rounded, so that only the very busiest places reach the top of each scale. Before publishing, the score was checked against a fixed list of well-known busy districts and quiet residential neighborhoods to make sure it put each group on the expected side.
A worked example of how to use it
Suppose two hotels you like fall in neighborhoods scored 82 and 51. The first sits in the “Quiet” band, the second in “Lively”. That does not settle anything on its own, but it tells you where to look next: for the second hotel, read recent reviews for mentions of street noise at night, check whether the room faces the main road, and see whether a quieter room can be requested. The score is best at telling you which questions to ask.
It is also useful the other way round. If you enjoy being in the middle of things, a low score near a station can simply mean restaurants and nightlife within a short walk.
Open data
The four ingredients and the final score for every neighborhood are published as quiet-derived.csv under the Open Database License (ODbL), because they include OpenStreetMap data. The complete formula is on the Data & methods page.
Open the map and pick “Estimated quietness” to see the pattern across the whole city in 3D, or read choosing an area to stay to put the score to work.