Supplemental Material — Functional Earned Value Trajectories for Construction Project Control: An FPCA Framework for Portfolio Diagnosis and Early Warning

# Supplemental Material — Functional Earned Value Trajectories for Construction Project Control: An FPCA Framework for Portfolio Diagnosis and Early Warning Authors: Joaquín Sancho Val, Carlos Cajal Hernando, Lourdes Martínez de Baños(Centro Universitario de la Defensa de Zaragoza, Academia General Militar, Zaragoza, Spain) Submitted to the ASCE *Journal of Construction Engineering and Management* (manuscript COENG-20531). ## Contents | File | Description ||---|---|| `Supplemental_Material.pdf` | Fig. S1 and Table S1 referenced in the paper || `FigS1.pdf`, `FigS1.png` | Fig. S1: smoothed functional trajectories of the derived EVM signals (CV/BAC, SV/BAC, CPI, SPI) || `TableS1_fpca_scores.csv` | Table S1: FPCA scores (PC1–PC3) of every project on every signal, plus cluster membership || `code/fdalib.py` | B-spline basis, penalized smoothing and FPCA (algorithms of the R package `fda`) || `code/fpca_evm_v26.py` | Full analysis pipeline: BAC normalization, FPCA, Ward clustering, stability checks, snapshot-PCA benchmark, leave-one-project-out early-warning experiment, figures || `code/replicate_v25.py` | Port of the earlier R (`fda`) settings, used to validate the Python implementation || `code/evm_monthly_summary.xlsx` | Processed monthly EVM panel (76 project-months, 8 projects) derived from the source dataset || `code/expected_output/` | Result tables produced by the pipeline (CSV and JSON), for checking a re-run | ## Reproducing the results ```pip install numpy scipy pandas openpyxl scikit-learn matplotlibcd codepython3 fpca_evm_v26.py # writes out/ (tables) and figs/ (figures)python3 replicate_v25.py # optional: reproduces the earlier R pipeline``` ## Source data The raw data are the openly available Project Portfolio Dataset:Thiele, B., Ryan, M., and Abbasi, A. (2020). *Project Portfolio Dataset*, version 2. figshare.https://doi.org/10.6084/m9.figshare.12998822.v2 (CC BY 4.0), described inThiele et al. (2021), *Data in Brief*, 34, 106659, https://doi.org/10.1016/j.dib.2020.106659 ## License Code and derived files: CC BY 4.0, consistent with the license of the source dataset.