• Building stone similarity to quarry rocks can be evaluated using in-situ measurements. • The similarity evaluation is made based on a specific chemical stone fingerprint. • Chemical fingerprint is developed using multivariate statistical analysis. • SIMCA and NBC are appropriate to evaluate the building-to-quarry stone similarity. The similarity of quarry stone with rapakivi stone of two buildings was assessed based on a 9-element fingerprint. The aim of the present case study was to assess the similarity between stones in buildings: the Kotka City Hall and the Orthodox Church of Kotka, and rocks in three Finnish quarries: Heponiemi, Jumalniemi and Metsola. Similarity assessment of building stones is necessary for effective and aesthetically-appropriate reconstruction. In the case of old buildings with stones of unknown origin, data on stone source quarries may have been lost over time. The present tool proposes a methodology to determine the most similar-to-the-origin stone and quarry for use in cultural heritage restoration . The tool in this study utilizes in-situ chemical analyses of the rock surface measured using a portable X-ray fluorescence (XRF) instrument together with modelling of the stone using multivariate statistical approaches (modified Soft Independent Modelling by Class Analogy , SIMCA) and machine learning ( Naive Bayes Classification , NBC) to determine the fingerprint of the stone in the building and assess its similarity to quarry rock data. The proposed methodology for identification of stone origin was utilised in a case study of rapakivi from the Kotka area in Finland.