Database for cross‑study buildability prediction in 3D printed concrete

This dataset supports the manuscript "Predicting Buildability of 3D Printed Concrete: A Cross-Study Machine Learning and Survival Analysis Framework" accepted for publication in Automation in Construction.

It contains 188 individual observations compiled from 26 independent studies published between 2018 and 2026, covering hollow cylinder and stacked layer geometries, plain and fibre‑reinforced mixes, and a wide range of static yield stress values (65–6910 Pa), print speeds (6–200 mm/s), layer heights (5–30 mm), nozzle hydraulic diameters (5–35 mm), and geometric factor α (0.036–2.518).

The dataset includes all necessary variables to replicate the survival analysis, machine learning classification (Extra Trees, Random Forest, etc.), Buildability Index (BI) calculations, and the design charts presented in the paper. Missing values (e.g., structuration rate A_thix, plastic viscosity) are left blank where not reported in the source studies.

Rows with incomplete critical data (e.g., missing layer width or yield stress) are flagged in the "remarks" column and were excluded from the primary analyses in the paper.

The file is provided as a CSV (UTF‑8) together with a README.txt that explains all columns, the formulas used to compute the geometric factor α, and the licensing terms (CC BY 4.0).

If you use this dataset, please cite both this Zenodo record and the associated journal article (full citation will be added after publication).