The Trexyz project: smart findability

All around us there are countless interesting collections — archaeological finds, historical objects, local nature data, volunteer‑assembled collections on all kinds of themes, and so on — which often exist only as spreadsheets or simple databases. They contain valuable information, but the institutions or individuals who manage them usually lack the resources, knowledge or infrastructure to present these data online in a modern, searchable way.

The Trexyz project explores how such hard‑to‑search datasets can be transformed in an accessible way into a rich, connected network that is suitable for online exploration and smart findability.

How does it work?

In the Trexyz approach, a table — a spreadsheet or an export from a database — is split, via its rows and columns and with the help of user‑friendly software modules, into three types of components that together form a reference network that can be used easily for professional‑level findability:

• The rows of the original table are converted into records, each describing a single object (for example a mill, an archaeological find or a chicken breed). This object record is structured in such a way that it works well in a simple yet effective findability system.
• The column values that refer to general, reusable concepts (such as mill type, period, colour) are grouped per type into lists. Where possible, these lists are given a hierarchical structure, which further improves findability. These concept lists form the semantic layer of the network.
• The combinations of concepts and the type of property they are linked to are stored as properties — these are the names of the property type combined with their specific values (such as location = Groningen, period = Mesolithic, colour = pearl grey). They connect object records with concepts and vice versa, forming the links between all parts of the network.

In the resulting reference network you can move freely: from an object to a concept, from a concept to another list, or via properties to a new selection. It forms the “engine” behind smart findability — comparable to faceted search, but more flexible and easier to implement.

Findability

On a website that uses the reference network, you can navigate the data — a collection, a product overview, research data, or anything else — in three ways:

• via the object records
• via the concepts
• via the properties, the most powerful way to search in a targeted manner

Search queries are built up step by step and stored in a search script. You can combine steps using AND, OR and NOT, or — for collections with coordinates — include spatial selections on the map.

By combining properties, concepts and search logic, relationships between objects become visible — whether they are Iron Age settlements, tower mills in Friesland or chicken breeds with a specific colour pattern.

Trexyz shows how a single generic model can be applied across very different domains, and how collections that previously existed only as spreadsheets can be transformed into a rich, searchable network that reveals connections that would otherwise remain hidden.

Who can create it?

The source of a reference network is the table in which each object — tangible or virtual — of the collection to be made findable is described in a single row. Based on this table, all used concepts are grouped per type into lists. These lists can be combined with existing lists and, where possible, are arranged hierarchically for optimal findability. The reference network is then created from the table and the concept lists. No technical knowledge is required: it is purely about understanding your own data and having ideas about how you want it to be findable. You can easily experiment to arrive at the optimal solution. The software handles the more technical work.

Examples

The reference networks and the info page in these examples are not translated yet.

Historic mills
A reference network of monumental mills, built from mill types, functions, locations and construction history, including support for spatial selection. Based on data from the Cultural Heritage Agency of the Netherlands, combined with publicly available web data.
View: Historic mills

Archaeological finds in Drenthe
A reference network based on archaeological finds from Drenthe, with periods, cultures, complexes and artefact types as meaningful entry points.
View: Archaeological finds

Dutch chicken breeds
A network centred on chicken breeds, with properties such as colour pattern, size, behaviour, comb type, weight classes, and more — showing how even small datasets can be made richly searchable.
View: Chicken breeds