Integrating Traditional Craftsmanship and Digital Fabrication for Heritage Restoration: A Dual-Site Comparative Framework from Lawang Sewu and Sam Poo Kong, Semarang

This dataset comprises the complete supplementary materials for a hybrid heritage restoration methodology integrating traditional craftsmanship with digital fabrication technologies (photogrammetry, structured-light scanning, and CNC milling). The materials document a dual-site comparative study conducted on two architecturally contrasting heritage structures in tropical Southeast Asia: a colonial-era Neo-Indo-European railway building with repetitive low-relief modular ornamentation, and a Sino-Javanese temple complex featuring high-relief figurative statuary and polychrome decoration. The dataset includes 17 discrete supplementary packages (S1–S17) plus a theoretical appendix (A), organised as follows: Theoretical and Methodological Documentation Full texts of the Venice Charter (1964) and Nara Document on Authenticity (1994) with custom interpretive guidelines for tropical heritage application Phase-by-phase restoration workflow protocols with quality checklists covering diagnostics, digital capture, CAD optimisation, CNC fabrication, artisanal finishing, in-situ installation, and post-restoration evaluation Detailed capture protocols for terrestrial short-range photogrammetry and structured-light scanning, including equipment specifications, parameter tables, and overlap strategies Site and Historical Documentation Comprehensive site records for both case studies: construction histories, architectural attributes, condition assessment tables with Damage Severity Index (DSI) calculations High-resolution panoramic source images (TIFF, 8,256 × 5,504 pixels) used for the article's comparative site map Digital Modeling and Fabrication Data Complete digital modelling pipeline: software specifications (Agisoft Metashape, RealityCapture, Rhinoceros 8, Grasshopper), step-by-step processing protocols, and Python code for Curvature Complexity Index (CCI) calculation CNC toolpath files (.NC format) and milling parameter tables: spindle speeds (20,000–24,000 RPM), feed rates (2.8–3.5 mm/s), step-over (0.08–0.12 mm), and estimated machining times Complete point cloud datasets (.PLY, .LAS) with quality reports: point densities (14–24 million pts/m²), registration RMSE (0.42–0.52 mm), stray-point and missing-data percentages Milling logs and surface roughness measurements (Ra 18–25 μm) with dimensional verification against Mitutoyo caliper references Evaluation and Validation Data Three-dimensional deviation maps (heatmaps in .OBJ, .STL, .PLY formats) with statistical summaries: mean absolute deviations (0.38–1.48 mm), percentage exceeding tolerance thresholds, and correlation analysis between DSI and deviation (Pearson's r = 0.62, p < 0.001) Complete expert panel evaluation protocols (7 independent experts, controlled conditions: 500 lux, 45° raking light, 1.5 m viewing distance, 48-hour test-retest) with raw scores across five criteria (visual authenticity, stylistic coherence, surface texture, colour matching, overall quality) and Krippendorff's α = 0.74 Public user survey instrument with 6 questions and raw response data (n = 50) Time-motion study dataset: traditional carving vs. hybrid times for 15 individual components across 5 typologies, with learning curve analysis showing 38.9% efficiency improvement Environmental and Qualitative Data Environmental monitoring data (temperature, relative humidity, solar radiation, rainfall, wind speed) across the study period Material compatibility test results: capillary absorption, porosity, and compressive strength for original vs. CNC-fabricated stone samples (all differences < 5%) Artisan interview transcripts documenting master craftsmen perspectives on digital workflow integration and authenticity Comparative and Analytical Frameworks Systematic comparative case study documentation against three international projects: Notre-Dame de Paris (CNC-assisted restoration), Hōryū-ji Temple (digital-artisan integration), and Milan Cathedral (HBIM for restoration management) Complete CCI dataset for 48 components with regression analyses: CCI vs. point cloud density (r = 0.94), CCI vs. absolute deviation (r = 0.81), and CCI vs. manual finishing time (r = 0.82), all with p < 0.001 Institutional permits and informed consent documentation File Formats and Technical SpecificationsFormats: PDF (textual documentation), TIFF (high-resolution images), PLY/LAS/OBJ/STL (3D data), NC (CNC toolpaths), XLSX/CSV (tabular data), GHX (Grasshopper definitions), PY (Python scripts). Total dataset size: approximately 15–25 GB depending on point cloud density and image resolution. License: Creative Commons Attribution 4.0 International (CC BY 4.0). Users are free to share, adapt, and build upon the material for any purpose, provided appropriate credit is given. Date of Collection: February–August 2025 Geographic Coverage: Semarang, Central Java, Indonesia (Lawang Sewu and Sam Poo Kong Temple) Language: Documentation primarily in English; permits and consent forms provided in both English and Indonesian. Keywords: Heritage conservation; digital fabrication; photogrammetry; CNC milling; traditional craftsmanship; authenticity; tropical heritage; Curvature Complexity Index (CCI); dual-site comparative framework; 3D point clouds; HBIM