Datasetpublic

gojiberries/softverse

Softverse: software referenced in social science replication code Which R, Python and Stata packages the code in published replication deposits loads, at economics and political science journals whose data-and-code policy an editor verifies. One row per deposit, file, file text, package reference, stated version and environment signal, plus per-language tallies. A reference is not a run. A row says deposited code names a package: in library(), an import, a Stata command. It does… See the full description on the dataset page: https://huggingface.co/datasets/gojiberries/softverse.

sourceHugging Faceotherupdated 7d agoView on Hugging Face

0likes130downloads

Dataset Card

Softverse: software referenced in social science replication code

Which R, Python and Stata packages the code in published replication deposits loads, at economics and political science journals whose data-and-code policy an editor verifies. One row per deposit, file, file text, package reference, stated version and environment signal, plus per-language tallies.

A reference is not a run. A row says deposited code names a package: in library(), an import, a Stata command. It does not say the code executed, and code an author kept out of the deposit cannot be seen.

Tables

config	rows	one row is
deposits	16,320	one replication deposit: DOI, journal, year, license, gaps
files	447,289	one file in a deposit: path, language, sha256, packages it loads
contents	277,543	one distinct file text, keyed by sha256 (open licenses only)
file_packages	366,525	one package a file references, with functions and pins
mentions	14,853,082	one reference in code: construct, line, resolution
package_versions	33,974	one stated version: manifest or install call
environment	19,983	one R/Python/Stata/Julia version or OS a deposit states

Per-language tallies (tally_by_function, tally_by_journal, tally_by_year, tally_python, tally_r, tally_stata) give, per package, the deposits that use it and the denominator: deposits with analyzable code in that language.

Coverage

source	deposits	with analyzable code	text published
dataverse	14,674	11,887	13,478
zenodo	1,646	1,358	1,621

Most-used packages:

—python: pandas (732), numpy (684), matplotlib (475), scipy (335), scikit-learn (271)
—r: ggplot2 (3,252), dplyr (3,134), tidyverse (2,398), stargazer (2,035), foreign (1,724)
—stata: estout (3,835), coefplot (1,671), outreg2 (1,553), reghdfe (1,535), ivreg2 (494)

Licenses

File text (contents) and code snippets (mentions.snippet) are published only for deposits whose license allows redistribution. Everything derived -- which packages a file loads, at which version -- is published for every deposit. Custom terms of use and unrecognised identifiers are treated as not redistributable. Each deposit's own license is in deposits.license_id; NC and ND licenses are included and flagged there.

license_id	text published	deposits
CC0-1.0	yes	13,330
CC-BY-4.0	yes	1,598
custom	no	1,140
CC-BY-NC-4.0	yes	63
none	no	61
MIT	yes	48
CC-BY-NC-SA-4.0	yes	20
CC-BY-NC-ND-4.0	yes	17
other-open	no	13
CC-BY-SA-4.0	yes	12
other-at	no	5
BSD-3-Clause	yes	5
CC-BY-ND-4.0	yes	2
GPL-3.0-or-later	yes	2
other-closed	no	1
other-nc	no	1
Apache-2.0	yes	1
ISC	yes	1

Known gaps

—186 deposits hold code only inside tar, 7z or rar archives too large to download here; their files are not in the corpus.
—Deposits whose collection did not complete: not_found 56, partial 34, failed 1.
—371 files over 1 MB keep their metadata row but not their text.
—The eight AEA journals, which deposit on openICPSR, are not included.
—A package vendored into a deposit (an renv library, a shipped .ado, a CRAN package's source tree) is marked is_vendor and not counted as used.

Source

Built by softverse. Package lookup and badges: <https://recite.github.io/softverse/>.