How we decide what is trending

Every spot on the radar carries a number from 0 to 100. It is the whole product, so it is worth being precise about what it means and how it is made.

The short version: the score is an estimate of how many people are going to be there. Not how good the event is, not how much we like it, and definitely not how much anyone paid. A quiet, excellent gallery opening scores low. A sold-out stadium scores high. That is the intended behaviour, because the question the radar answers is "where is everyone going tonight", not "what would a critic recommend".

What the number is built from

We pull events from ticket sellers, sports fixture feeds, national and municipal open-data calendars, and public-holiday sources, then derive one score from the signals those feeds actually carry.

Crowd size is the spine. Where a listing carries a real expected attendance we use that; failing that, the venue's capacity; failing that, for a sports fixture, the level of the competition (a cup final draws differently from a midweek third-division game). A real measurement always beats an inference, and the largest real figure available wins.

Where we cannot measure the crowd, we estimate it from evidence, and we say so. An act booked for forty dates is a bigger operation than one playing a single night in a back room. A venue with twelve thousand Google reviews is a bigger room than one with forty. An act people are actively looking up right now is drawing more attention than one nobody is searching for. None of those is an attendance figure and none is treated as one. They separate the long tail so that "we do not know how big this is" does not collapse every unlisted show onto the same number.

Everything else only tilts. The kind of event, the venue's rating, how many independent sources list it: each nudges the score a little around the crowd estimate. The tilt is small by design, so an event that is genuinely much bigger still outranks a smaller one in a more fashionable category.

What is deliberately not in it

Money. No promoter, venue or ticket seller can pay to rank higher. There is no advertising inventory on the board, no promoted slot, and no affiliate arrangement that would reward us for pushing one listing over another. If that ever changes it will be labelled on the card itself, not buried here.

Time. How soon an event starts is not part of the score. Sorting by time is a separate control, so folding it into the number would double-count it and make the ranking fight the sort.

Our taste. There is no editorial thumb on the scale, and no source gets a bonus for being a source. One model runs over everything. What does differ is how much each feed tells us: a fixture list publishes the competition, a ticket seller publishes the venue, and an open-data calendar often publishes neither, so a well-described event can be sized more confidently than a sparsely described one.

What it cannot tell you

The score is only as good as what the feeds publish, and they are uneven. Ticketed events are covered far better than free ones. Club nights and community events are systematically under-represented, because the places that host them often do not publish a machine-readable listing at all. A city with a strong open-data programme looks busier than a comparable city without one. That is a coverage gap, not a judgement about the city, and closing it is most of the work.

Some events carry a date but no start time, because the promoter has not announced one yet. We place those at midday local so they can never drift onto the wrong day, but the card still shows a time we chose rather than one anyone published. Labelling them properly is on the list.

Why do it this way at all

Because every other way of ranking a city's nightlife has an owner with a stake in the answer. Ticket sellers rank what they sell. Venue guides rank who advertises. Social feeds rank whatever holds attention longest.

A neutral ranking across every source is only useful if it is genuinely neutral, and neutrality is not a claim you can make credibly without showing your work. So: the number is a crowd estimate, it is derived from the signals above, and nothing else touches it.