Replication package for "Counts versus kilowatts: city population and the configuration of public charging in China"

This replication package accompanies the manuscript "Counts versus kilowatts: city population and the configuration of public charging in China" by Ziran Kong, Dongwei Tian, Yanyun Mao and Jian Zhuo, submitted to Communications in Transportation Research. The study is an exploratory cross-section of 336 Chinese cities. It splits the population exponent of listed public charging capacity into stations, connectors per station and rated power per connector, and compares cities at similar numbers of new energy vehicles (NEVs) per resident. Contentsdata/city_summary.csv: one row for each of the 336 cities, with station, connector and listed-capacity summaries from the charging layer of a commercial web map platform (snapshot of 11 November 2025), census population of 2020, alternative population layers, and registered NEV stock at the end of 2021, 2022 and 2023. data/city_summary_codebook.csv defines each column.data/provincial_nev_official.csv: official provincial NEV stocks with their source sentences and addresses.data/Table_S65_literature.xlsx: the literature table of the Supplementary Material (166 studies). registry/: every number reported in the main text and the Supplementary Material, with its source script. reproducibility_map.csv: for each table and figure, whether it is rerun from this package and, if not, what is missing. rerun/: scripts that rerun the main results from the city-level table and compare them with the registry. 40_main/, 41_verify/ and src/: the analysis source code listed in Table S64, kept as a record of the methods. README.md, LICENSE_NOTE.md and the journal's replication explanatory file. How to rerunFrom the package root, run "Rscript --vanilla rerun/run_main.R" and then "python rerun/plot_figures.py". The first script needs R with clubSandwich and jsonlite; the second needs Python with numpy, pandas and matplotlib. Together they rerun Tables 2, 3, S18, S19, S21, S25 and S26, Figs. 5, 6 and 9 (redrawn in a plain style) and parts of Tables S27, S35 and S39, without bootstrap intervals, and compare 332 values with the registry. The scripts do not depend on the system locale. LimitsStation-level records cannot be redistributed, so the package holds city-level summaries. Results that use station-level records, and the bootstrap intervals, cannot be rerun from the package. The analysis source code is a record of the methods and does not run as a stand-alone pipeline. LicencesData and documentation: CC BY 4.0. Code: MIT. Quoted passages in data/provincial_nev_official.csv and data/Table_S65_literature.xlsx remain under the terms of their sources.