This repository publishes a tracked annual release snapshot of the
fairadi-data
project. Provenance for a release is established by the combination of:
- the git release reference and repository history
MANIFEST.tsv, which records tracked file paths, sizes, and SHA-256 digests- the build scripts under
Makefile,ACS5/,Decennial/,FIPS/,ADI/,CDI/, andutilities/ - release metadata in
metadata.json,CITATION.cff, andro-crate-metadata.json - the DCAT-US catalog record in
dcat-us.json - the formal PROV-N serialization in
provenance.provn - the split-license statements in
LICENSEandLICENSE-data
MANIFEST.tsv applies to the tracked repository release snapshot. It does not
describe the tarballs emitted for Zenodo upload. The Zenodo packaging workflow
produces a separate SHA256SUMS file for those packaged release artifacts.
The Zenodo DOI for the current release record is:
10.5281/zenodo.19222629
The intended git release reference for this release is:
v1.0.0
The canonical released ADI dataset artifact is:
ADI/fairadi.csv.gz
The column definitions for this file are documented in:
ADI/fairadi_data_dictionary.tsvADI/fairadi_schema.jsonADI/fairadi_codelists.tsv
The canonical released CDI dataset artifact is:
CDI/faircdi.csv.gz
The column definitions and row schema for this file are documented in:
CDI/faircdi_data_dictionary.tsvCDI/faircdi_schema.json
The CDI process documentation is:
CDI/README.md
The public-data workflow draws from:
- U.S. Census Bureau American Community Survey 5-year tables in
ACS5/ - U.S. Census Bureau Decennial Census tables in
Decennial/ - Census geography inventories and supporting reference files in
FIPS/
For comparison diagnostics in ADI/README.Rmd, the project also references
Neighborhood Atlas files that are not redistributed in this repository.
At a high level:
FIPS/builds geography inventories.ACS5/andDecennial/build the tracked Census table extracts.ADI/builds topic-level derived files, suppression inputs, and the finalfairadi.csv.gzdataset.CDI/builds the total-population / housing-unit replacement flags, the 18 CDI components, and the currentfaircdi.csv.gzoutput.MANIFEST.tsvinventories the tracked release contents.
The top-level workflow is orchestrated by:
Makefile
The final dataset builds are orchestrated by:
ADI/MakefileADI/fairadi.RCDI/MakefileCDI/faircdi.R
The repository also publishes a formal provenance serialization in:
provenance.provn
This file uses the W3C PROV-N notation to describe the release snapshot, selected generated artifacts, principal build activities, creators, and core derivation relationships. It is a concise machine-readable complement to this human-oriented provenance document, not an exhaustive file-by-file execution log.
Validate the provenance file with:
make validate-provenanceThis runs utilities/validate_provenance.py, which performs repository-level
consistency checks. If provconvert happens to be installed locally, the
script also uses it as an optional extra parse check, but it is not required
for the repository workflow.
Validate the RO-Crate metadata with:
make validate-ro-crateThis runs utilities/validate_ro_crate.py, which performs repository-level
checks on ro-crate-metadata.json, including the required root entities,
selected release resources, action entries, and referenced local paths.
To refresh the tracked-file manifest:
make manifestThis refreshes the repository-snapshot manifest, not the Zenodo package checksums.
To verify any single tracked file against the manifest:
awk -F '\t' '$1 == "ADI/fairadi.csv.gz" {print $4 " " $1}' MANIFEST.tsv | shasum -a 256 -cTo verify all tracked files listed in the manifest:
python3 - <<'PY'
import csv
with open("MANIFEST.tsv", newline="", encoding="utf-8") as f:
r = csv.DictReader(f, dialect="excel-tab")
for row in r:
print(f"{row['sha256']} {row['path']}")
PYPipe that output to shasum -a 256 -c if you want a full integrity check.