Research / Projects / Data We Kept

Data we kept.

Public datasets DaedalMap hosts after their publishers stopped distributing them. Mirror provenance, drift notes, and source status carried into pack attribution.

Preserved sources

Three packs draw from preservation mirrors after their original publishers discontinued distribution between January 2025 and February 2026. Methodology is unchanged. Coverage and schemas match the last canonical release.

Preservation criteria

A pack carries the preserved status when all of the following hold:

  • The original publisher has stopped distributing the dataset.
  • A community or institutional mirror exists and has been retrieved.
  • The mirrored copy has been hash-verified or row-compared against the last known canonical release before discontinuation.
  • Any value drift between mirror and last-known-original is documented.

Reference data updates on five-to-ten-year cycles. A missed annual refresh leaves the data useful for years. When a publisher resumes distribution, the source status returns to active and the pack continues unchanged.

CIA World Factbook

  • Pack: world_factbook
  • Original publisher: Central Intelligence Agency
  • Discontinued: February 4, 2026
  • Stated reason: None given by the publisher
  • Mirrors used: Mozilla Data Collective, Factbook Archive
  • Coverage: 1990 - 2025, 111 indicators across 195+ countries
  • Normal cadence: Continuous on the publisher's site; annual snapshots commonly cited in research

FEMA Future National Risk Index

  • Pack: future_nri (in preparation)
  • Original publisher: Federal Emergency Management Agency
  • Removed: February 2025
  • Stated reason: Not publicly disclosed; subject of a 2025 lawsuit dismissed for lack of standing in March 2026
  • Mirrors used: EELP Harvard Law tracker; fulton-ring/nri-future-risk recreation repository
  • Coverage: Single Dec 2024 prototype release; county scale; five hazard families across mid-century and late-century, lower and higher warming scenarios
  • Mirror enrichment: The fulton-ring recreation carries ten additional columns the original FEMA file omitted, including the Coastal Flooding hazard family. Methodology source is the EELP-hosted technical document.

Climate and Economic Justice Screening Tool (CEJST)

  • Pack: cejst (in preparation)
  • Original publisher: Council on Environmental Quality, White House
  • Removed: January 22, 2025
  • Stated reason: Public access discontinued by the publisher; the tool and underlying data remain mirrored by community partners
  • Mirrors used: Public Environmental Data Partners CloudFront CDN; EDGI gov-data archiving
  • Coverage: v2.0 release (Dec 2024), 74,134 census tracts, 136 columns
  • Mirror drift: The PEDP mirror differs from the last DaedalMap-side canonical copy in 216 rows across three columns. Plausible explanation is a late publisher patch captured by the mirror after the canonical copy was retrieved. The mirror is treated as canonical going forward.

Preservation in the operating layer

Public reference data is load-bearing for research. The pipeline that keeps it queryable carries the same maintenance discipline as freshness for live data: mirror selection, hash verification, drift tracking, and provenance recorded against every pack. The list above grows as additional sources move into the preserved status.