Dataset and analysis code for: "The price of uncertainty in perovskite photovoltaics: a Bayesian levelized cost analysis"
Description
Contents: 1. device_dataset.csv — the consolidated device dataset: the seventeen perovskite, tin halide and tandem architectures with their simulated electrical outputs (efficiency, Voc, Jsc, fill factor) as reported by the peer reviewed source simulation studies cited in the article. These are idealized simulation results, not measured devices; certified record efficiencies used as the article's conservative baseline are in the workbook sheet certified_benchmarks. 2. perovskite_TEA_inputs_REAL.xlsx — the input data workbook (12 sheets): climate scenario priors (climate_scenarios), cost priors (cost_priors), material resolved bill of materials (bill_of_materials, material_prices), hole transport layer substitution inputs (htl_substitution), temperature coefficients (temp_coefficient), process quality inputs (defect_quality), device efficiencies (device_efficiencies), source mapping (sources), certified records (certified_benchmarks) and economic assumptions (economics_inputs). Each sheet lists the distribution and the source of every input; see also Table A.1 of the article. 3. python_code.docx — the complete, documented analysis notebook: Monte Carlo levelized cost model, hierarchical Bayesian pooling (PyMC), Sobol sensitivity analysis, value of information, financing scenarios and real options valuation, including the code that generates every figure of the article.
Files
Steps to reproduce
The code runs in a free Google Colab environment (Python 3) with numpy, pandas, matplotlib, scipy and PyMC; no proprietary software is required. Copy the cells of python_code.docx into a notebook in order and run top to bottom. Random seeds are fixed in the code, so all reported values reproduce exactly.