Archaeological and geoscientific research often relies on spatial data whose precision and reliability vary considerably. Coordinates derived from excavation reports, literature, or modern surveys frequently embody vagueness, ambiguity, or conflicting interpretations. This paper addresses such wobbling geospatial data by presenting a semantic approach that embeds uncertainty directly into Linked Open Data (LOD). At the core of this approach lies the Fuzzy Spatial Locations Ontology (FSL-O), which builds on PROV-O, SKOS, and GeoSPARQL to model entities, activities, and agents alongside explicit properties for certainty, provenance, and methodological detail. Its implementation in the fuzzy-sl Wikibase extends this framework through a dedicated schema for geolocation and coordinate metadata, enabling the documentation of precision, methods, sources, and actors. The SPARQL Unicorn Toolkit further supports visualisation by integrating these data into GIS environments and generating human-readable HTML outputs. The methodology is demonstrated through interdisciplinary case studies: Ogham stones with multiple biographies between ancient findspots and museum contexts; Campanian Ignimbrite deposits serving as Late Pleistocene chronological markers; poorly documented numismatic hoards from Croton; and Roman Samian ware distributions. Each example illustrates how uncertainty can be expressed as structured, machine-readable information, rather than being hidden or oversimplified. By embedding fuzziness into LOD, this work enhances the FAIRness, transparency, and interoperability of archaeological and geoscientific datasets. It contributes to international debates on how to responsibly represent spatial uncertainty and provides a foundation for future extensions of ontology and Wikibase models across broader cultural heritage and scientific domains.