Flows

The Marquenterre Park, in the Somme Bay, is testing an artificial intelligence tool capable of forecasting its attendance up to six months in advance.

The objective is to better anticipate peak crowds and optimize the site’s organization, while preserving this protected natural space.



With nearly 160,000 visitors welcomed each year, the Marquenterre Park is among the most visited natural sites in Hauts-de-France.

This appeal is nonetheless accompanied by peaks in attendance concentrated on certain days and time slots.

In this protected natural area, where the welcome is limited to 300 visitors per hour, these periods of high traffic can strain the conditions of reception, the teams and the various visitor services offered.

Until now, managing these peaks relied mainly on the observation and experience of the on-site teams.

Developed as part of an experiment supported by the Destination France plan, the model is able to forecast attendance levels up to six months ahead, with forecasts broken down by hourly intervals.

To establish these projections, the tool crosses several types of data likely to influence tourist behavior: French and Belgian school holidays, public holidays, weather conditions, events held in the area or even snowfall in the mountain ranges, which can influence destination choices.



These data are then gathered into a dashboard for the park teams. This “pilot cockpit” allows visualizing expected crowd levels and their evolution over time.

The artificial intelligence does not directly make operational decisions. It provides teams with a forward-looking view of attendance to help them adjust their organization.

To read : Weekend of disconnect in the Bay of Somme

This can help better organize human resources, prepare the dining services and mobilize more resources during peaks of attendance.

“For as long as the Marquenterre Park has sought to find the right balance between tourist attractiveness and the preservation of its natural heritage, with this experiment we are taking a new step.

AI becomes a tool for the territory, able to help managers better understand visitor flows and anticipate their needs“, explains Emmanuel Maquet, president of the Destination Baie de Somme management.

The Marquenterre Park currently serves as the testing ground for this system. But the stated ambition goes beyond the site’s boundaries.

The project is among the 13 initiatives selected from 75 applications under the Destination France plan.

In the long term, the model could be enriched with new data to better understand visitors’ movements at the scale of the Bay of Somme.