Measuring Agricultural Reconstruction under Statistical System Failure

Open Science Replication Package

This repository contains the complete replication package accompanying the manuscript "Measuring Agricultural Reconstruction under Statistical System Failure: Evidence from Post-Conflict Azerbaijan".

---

Repository Overview

This repository follows open science and computational reproducibility principles and includes:

• Complete Python source code
• Raw satellite and administrative datasets
• Reproducible computational workflow
• Comprehensive documentation
• Software environment specifications
• Replication guide

---

Repository Structure

AJAE_Replication/
├── code/
├── data/
├── docs/
├── manuscript/
├── output/
├── figures/
├── gee_notebook/
├── README.md
├── CHANGELOG.md
├── CITATION.cff
├── LICENSE
├── requirements.txt
├── environment.yml
└── .gitignore

Documentation

• docs/CODEBOOK.md -- Analytical workflow and script-by-script description
• docs/DATA_DESCRIPTION.md -- Data sources and dataset structure
• docs/REPRODUCIBILITY_CHECKLIST.md -- Reproducibility checklist
• docs/Replication_Guide.md -- Complete replication guide

Installation

conda env create -f environment.yml
conda activate ajae-repro

or

pip install -r requirements.txt

Run

python code/run_all.py

This reproduces the complete analytical workflow: synthetic-control estimation, randomization and placebo inference, external validation against official statistics, cross-validation against independent satellite indicators, the event-study specification, seven robustness and sensitivity analyses (Appendix C), the national extensive/intensive margin decomposition (Appendix A), and all manuscript tables and figures.

Expected runtime is under three minutes on a standard laptop.

Script-to-Manuscript Correspondence

Script	Produces	Manuscript location
01_table1_donor_weights.py	Table 1	Section 4.2
02_figure2_ndvi_trajectories.py	Figure 2	Section 5.1
03_figure3_reconstruction_effects.py	Figure 3	Section 5.2
04_table2_permutation_inference.py	Table 2	Section 5.3
05_figure4_statistical_validation.py	Figure 4	Section 5.3
06_table3_external_validation.py	Table 3	Section 5.4
07_figure5_crossvalidation_scatter.py	Figure 5	Section 5.5
08_figure6_satellite_imagery.py	Figure 6	Section 5.5
09_figure7_event_study.py	Figure 7	Section 5.6
10_figureB1_district_map.py	Figure B1	Appendix B
11_tableC1_pretreatment_coefficients.py	Table C1	Appendix C.2
12_appendixA_national_decomposition.py	Appendix A figures	Appendix A
13_appendixC1_leave_one_out.py	Leave-one-out robustness	Appendix C.1
14_appendixC2_cluster_robust.py	Cluster-robust event study	Appendix C.2
15_appendixC2_joint_pretrend_permutation.py	OLS + permutation joint test	Appendix C.2
16_appendixC3_placebo_in_time.py	Placebo-in-time test	Appendix C.3
17_appendixC4_ridge_synthetic_control.py	Ridge-regularized SCM	Appendix C.4
18_appendixC5_power_analysis.py	Minimum detectable effect	Appendix C.5
19_appendixC6_staggered_heterogeneity.py	Timing heterogeneity	Appendix C.6
20_appendixC7_qa_ndvi_check.py	QA-mask sensitivity check	Appendix C.7
21_figure1_conceptual_framework.py	Figure 1	Section 2.4

Citation

Please cite both the published article and the archived GitHub/Zenodo repository. Citation metadata are provided in CITATION.cff.

License

MIT License (code). Data files retain the licensing terms of their original sources; see docs/DATA_DESCRIPTION.md.

Contact

Halil Tosun, Ph.D.

ORCID: https://orcid.org/0000-0001-5117-0390

Email: halilibrahimtosun@gmail.com

Zenodo DOI: https://doi.org/10.5281/zenodo.21435733

Version: 1.0.0