Staffing plans for every entrance, based on arrival forecasts

Event Staffing Planner forecasts guest arrivals for each entrance and hour, shows the uncertainty in each forecast, and proposes staffing plans. The final decision stays with the planner.

This introduction takes about one minute.

The planning problem

Every event requires a decision on how many staff to assign to each entrance, hour by hour.

Understaffing

Long queues, lower guest satisfaction and guests arriving late for the show.

Overstaffing

Idle staff and labour costs that could be used elsewhere.

Attendance is uncertain. Weather, competing events nearby and changes to access, such as a new car park, all affect arrivals.

How it works

Forecasts arrivals
For each entrance and hour, a median estimate and an 80% range.
Incorporates local knowledge
Roadworks, nearby events and schedule changes can be added as constraints. They apply only after you confirm them.
Proposes staffing plans
A range of plans from lowest cost to shortest queues. You choose the trade-off.

It is available on this website and in Claude through an MCP connector.

About this prototype

This is a research prototype that uses synthetic data.

Synthetic data

Parkland Theme Park, its three entrances and 120 days of visitor history are fictional. The data reproduces realistic arrival patterns: morning peaks, an evening rush before the show, rainy days and a newly opened entrance.

Computed results

No figures are entered by hand. All forecasts and plans are computed from this history by a forecasting model and an optimization algorithm.

Validated forecast

The forecast's 80% ranges are checked against actual arrivals on days held out from training.

Simplified methods

The model and the algorithm are intentionally simple, designed to test the research questions. A production deployment would use the venue's own data and stronger models.

Research question: whether explaining the sources of forecast uncertainty, and converting planners' knowledge into constraints for the optimizer, leads to better staffing decisions.

Get started

Sign in to receive a private copy of the demo park. Your workspace is visible only to you and can be reset at any time.

Sign in

or

To use it in Claude, add a custom connector with this address, then ask, for example, Show me the forecast for Halloween Night.

Custom connectors require a paid Claude plan. How your data is handled