Twelve months of 68,930 club nights, counted three ways. Where the floor actually fills, when a city is having its moment, and what the room is like once you are inside it.
London, Berlin, New York and Barcelona hold the top four places in this dataset at every threshold tested, without a single swap, in all twelve months. That is a robust finding and a useless recommendation.
It is useless because it answers a question nobody asks. Nobody books a weekend around which city has the most club nights in aggregate. They ask where to go in March, whether the room will have three hundred people in it or thirty, and whether the door will let them in. A ranking of totals cannot answer any of that, and the cities at the top of it are precisely the ones where the answer barely changes month to month.
So this piece does two things. First it finds the moments: the specific city-months where a scene visibly lifts, using a test strict enough to throw out most of what looks like a spike. Then it puts those moments next to the thing the numbers cannot see, which is what the night is actually like. A 500-RSVP night in Detroit in May and a 500-RSVP night in Tokyo in May are not the same event, in the same room, under the same rules, with the same relationship to a phone camera.
Counting events at one cutoff makes spikes easy to fake. A promoter dumping fifty listings inflates the 20+ count without a single extra person dancing. One festival landing inflates the 500+ count without the city's ordinary weekends changing at all. Neither is a scene doing something.
So the test here is concordance: a city-month only counts as a surge when it lifts at all three cutoffs at once, measured against that city's own median for the other eleven months. More nights, bigger nights and headline nights, together. The score is the weakest of the three lifts, so a month is only as strong as its softest link.
The two tables below are the false positives a single-threshold count would have swallowed. On the left, months where listings jumped but the big nights actually shrank. On the right, months where one enormous event landed on an otherwise ordinary calendar. Both look like a scene surging. Neither is one.
Paris in October 2025 is the cleanest example. Listings rose 50% over its own baseline while events over 500 RSVPs fell to 57% of normal. A naive count would have called that a boom. It was the opposite.
Six cities, twelve months, shape only
Events at the 20+ cutoff, each scaled to its own peak. Shape is comparable across cities even when rank is not.
Nine scenes, each with what the data says and what the data cannot say. The cultural claims here are sourced; every one traces to a citation in Appendix G. Where something is unexplained, it says so.
Most of what separates these scenes is not the music. It is the rules, and the rules exist for reasons that are usually older and more serious than they look from outside.
At Berghain in Berlin and Bassiani in Tbilisi, staff cover phone cameras on the way in and ejection follows if you peel it off. Berghain's is about protecting a space built for a queer crowd that wants to dance without being documented. Bassiani's is about a club that was raided by armed police in 2018 and has faced hostility from right-wing groups since. Treat the sticker as the price of entry, not an obstacle.
Berlin's clubs are not obliged to close, which is why a Saturday night is a Sunday afternoon. Amsterdam issues a small number of 24-hour licences, deliberately to venues outside the busy centre so crowds do not all hit the street at once. In Japan, dancing after midnight was effectively illegal until the entertainment law was amended in 2015. Arriving at midnight is early in one city and closing time in another.
Selective doors in Berlin are widely misread as arbitrary. The consistent thread in how these clubs describe themselves is protecting the crowd inside rather than filtering for glamour, which is the same logic as the camera stickers. Go in small numbers, know who is playing, and do not treat the queue as an obstacle to be negotiated.
Ibiza clears 100 RSVPs on only 6% of its listed events, not because its parties are small but because its clubs sell through their own channels rather than through listings platforms. Low numbers on a listings site can mean a small scene, a closed scene, or simply a scene that sells its tickets somewhere else. Do not read absence as evidence.
Resident Advisor is a listings platform with a footprint, not a census of world nightlife. Two blind spots matter enough to name before anyone quotes a rank.
The first is platform share, and it is now measured rather than assumed. A sample of 22 cities against Shotgun found that in Paris and Lyon, Shotgun carries the overwhelming majority of listed events, so Resident Advisor sees a minority of what is happening. In Berlin, London, Glasgow and Detroit, Shotgun is effectively absent and the Resident Advisor picture is close to complete. The correction is real, it is large, and it is confined to specific countries.
The second is Asia, and Shotgun does not explain it. Shotgun has no city page at all for Tokyo, Bangkok, Seoul or Melbourne, so it cannot be the missing platform there. Something else is. Japan's own listings ecosystem runs through iFLYER and clubberia, which announced a partnership in 2018 to share event data and ticketing. Thailand's runs through Ticketmelon and Eventpop, Indonesia's through Loket. None of them are in this dataset, which is why Tokyo appears in the data as a city with heavy listings and almost no crowd.
Shotgun versus Resident Advisor, share of currently listed events
22 cities, upcoming events on each platform, sampled 8 Sep 2026. Shotgun counts all genres, so its share is an upper bound.
The hard data behind every claim above. Tables are sortable by eye, charts are hoverable, and the underlying files are listed at the end.
attending field counts Resident Advisor users who marked themselves attending. It is the only per-event crowd figure the platform publishes. It is a fair proxy for scale and it is not verified attendance, and any article using it has to say so rather than imply door numbers.Events per month at each cutoff
Pale 20+, mid 100+, solid 500+ (figure below each month). Hover for detail.
Monthly events over 500 RSVPs, top three destinations
The only podium break in twelve months is June 2026, when Barcelona finished first.
Below fourth place the cutoff is the ranking. Detroit climbs from 15th to 6th as the bar rises, Amsterdam slides from 5th to 9th, and Tokyo leaves the chart.
Rank at each cutoff
Top 14 at 500+, plus Tokyo. Grey lines hold station.
The share of each city's 20+ events that also clear the higher bars. Globally 41% clear 100 and 7.7% clear 500. Where a city sits far below its peers, an absolute cutoff is measuring its RSVP culture as much as its parties.
Survival rate at both bars, the twenty biggest scenes
Dashed lines mark the global rates.
Currently listed upcoming events on each platform, sampled the same day. Shotgun's city pages publish their own count, which is what is used here. Two caveats: Shotgun lists all genres while Resident Advisor lists electronic only, so Shotgun's share is an upper bound; and an event listed on both platforms is counted twice, which also inflates it. Shotgun's per-genre pages were tested and rejected, because they report the city-wide total rather than a filtered count.
Rows flagged thin have fewer than 30 qualifying events across the whole year. Treat their exact position as noise.
Numbers come from the crawl. Everything about culture, law and policy above traces to one of these.