I have spent a year measuring the queues of twenty-seven European theme parks across ten countries.
What came out of it is four tables and a dossier. This page exists to say precisely what is in
them, how they were built, and what can be done with them.
The four tables
Parks (27 rows, 28 columns). Declared attendance with its source and year, operating days
counted on the official calendar, average daily attendance, number of rides with a wait-time
record, waits across the five and the ten worst queues, portfolio wait, three in-house queue
efficiency metrics, and site area where it is on record.
Attractions (438 records, 177 of them inside the twenty-seven parks). Manufacturer for 375,
model, trains, seats per train, cycle duration, declared capacity for 163, cost in its original
currency for 52, and height, speed and opening year for all of them.
Waits (1,197 rows). Mean wait and mean of daily maximum waits, ride by ride.
Cost per installed person-hour (40 rows). The declared cost of the ride brought to constant
2025 euros and divided by its declared capacity. I am not aware of another table that holds this.
How it was built
Waits come from the public historical record of posted wait times, a full year, ride by ride.
Technical records come from RCDB, opened one entry at a time. Attendance comes from the TEA/AECOM
Theme Index and from operators. Operating days were counted on each park’s published calendar.
Two filters change the result and you should know they are there. Single-rider rows are excluded,
because that is a different queue. And frozen feeds — rides posting the exact same value for months
— are dropped by a blind rule applied identically to all twenty-seven, rather than by hand park by
park. What I found cleaning my own data is written up here.
Costs are converted to euros using the irrevocable rates of the currencies that joined the euro and
the ECB annual rates for the rest, then deflated with harmonised HICP. Anything before 1996 is
flagged as an underestimate, because the price series does not reach further back.
What is missing
Site area is on record for only five of the twenty-seven parks. Cost, for 52 of the 438
attractions. Declared design peak, for almost none. The gaps are marked as gaps in the tables
rather than filled with estimates, which is what makes them auditable.
For professional use
What can be commissioned is analysis, not the data dump. That distinction isn’t commercial, it’s
about rights: the waits come from a third-party feed and the technical records from someone else’s
database, so what is mine, and what I can put in a contract, is what I have computed with them.
Three things, concretely:
A single-park report. Its queue efficiency measured against the norm of the twenty-seven, ride
by ride, with the cost per installed person-hour of what it owns and where its real bottleneck sits.
A comparative report across several parks in one chain, or one park against its direct
competitors.
A technical note on one specific attraction: its capacity ceiling and what that ceiling depends
on. It’s what I did in the open
with Bakunawa.
And if what you need is a figure these tables already hold, to quote in an article or a report, just
ask and I’ll send it with its source.
All of it by email: get in touch.
Getting the tables
The four tables as CSV and the full dossier — 222 pages, twenty-eight chapters, one chapter per
park plus the whole method — are available at the Engineer tier on
my Patreon.
The wait times behind this study are built on the data service at
Powered by Queue-Times.com. Attraction technical records come from
RCDB, and attendance figures from the TEA/AECOM Theme Index and from
operators themselves.
And if you would rather see how I work first, the analysis of
Bakunawa’s capacity
and the one on declared capacity by decade
both come out of these tables, and both are complete and sourced.