Case Study:
Havelland Tour Portal
A geodata-based tour portal for active tourists in the Havelland region
Background
An internal research platform for hiking and cycling routes in the Havelland region was to become the foundation for a public tour portal dedicated to the area. Unlike well-known outdoor portals such as Komoot or Outdooractive, the focus was not meant to be on tours alone. The goal was to also capture the geographic, landscape, cultural, and historical features of the Havelland and to relate them to one another.
Project
Havelland-Touren is aimed at active tourists who want to explore the region by bike, on foot, or on the water. Unlike classic tour portals, the focus is not on the individual tour alone. Instead, the portal models the region itself.
Alongside around 130 tours, the database contains about 70 bodies of water and 1,500 points of interest. Added to this are regions, nature reserves, biotope data, and specialized classes of places such as bathing spots, restaurants and cafés, historic buildings, windmills, and bird-watching sites.
The data model is deliberately oriented around the natural and cultural features of the Havelland. The Havel river and its many lakes and other waters shape large parts of the landscape. Bodies of water are therefore not just managed as geographic objects in the background but form their own information areas. Bathing spots, fishermen, sights, and tours all relate back to them in turn.
The result is an interconnected body of information in which a tour is just one of several possible entry points into the region. From a tour, a visitor might, for example, reach a body of water, discover a bathing spot there, and then come across another tour. The portal is thus meant not only to offer concrete tour suggestions but to create a digital space for exploration that invites visitors to browse and discover the Havelland.
Implementation
Havelland-Touren is technically built on Python and Django, with PostgreSQL and PostGIS for storing and processing the geodata. The application runs on a Kubernetes infrastructure.
Tours, bodies of water, biotopes, and other geographic objects are stored in the database with their actual geometries. This allows many relationships to be derived automatically from their spatial position. Only a tour’s start and end points are set editorially; which relevant geo-objects lie along a route is something the application determines on its own.
The geodata is not only used to display objects on a map. Through spatial calculations, the system gains additional information about the character of a tour. For example, a tour’s geometry is intersected with the areas of the available biotope data. This makes it possible to determine which landscape types characterize a route. Likewise, the application calculates over what distances a tour runs directly along bodies of water. If a route runs through forest for 70 percent of its length and along shorelines for eight kilometers, for instance, that already paints a fairly meaningful picture of its landscape character.
A particular challenge lies less in the technical linking of the data than in the question of which of these relationships are actually relevant to a user. PostGIS makes it possible to determine which objects intersect, which POIs lie near a route, or which tours pass a given body of water. A purely technical, omnidirectional linking of all geographically related objects, however, does not automatically lead to a good user experience.
That is why the semantic relationships between the different object classes are deliberately modeled. The technical relationship between two geo-objects is merely the foundation; only through domain-specific classification does it become a connection that is meaningful to the visitor.
The same applies to user guidance. Havelland-Touren is not meant to present the available data as exhaustively as possible on every page. Instead, editorial text guides visitors through the content, with the structured data and automatically derived relationships supporting these texts and opening up opportunities for further browsing.
Over the years, Havelland-Touren has thus become a practical testing ground for geodata-based information systems. Working with real tours, complex geometries, and an extensive set of diverse geo-objects has provided valuable experience in handling geodata and its reference models, as well as in optimizing spatial queries.
At the same time, the portal benefits from experience gained in other projects – from Django and modern frontend technologies to operation on Kubernetes. The insights gained here in turn feed back into the development of other information- and geodata-centric applications.
We can also implement your project!
Do you want to implement a similar project? Or do you have a completely different idea? Talk to us and we will find the best solution together.