Methods & Caveats

WGU Atlas is built from 108 public WGU catalog editions spanning January 2017 through March 2026. Three editions are absent from the archive (2017-02, 2017-04, 2017-06), likely never published as separate snapshots on the WGU public catalog page.

Each edition represents a distinct published snapshot of WGU's public course catalog. The parser extracts course codes, titles, program memberships, and structural metadata from each edition.

The WGU catalog underwent a structural formatting change in mid-2024. Two parser eras are recognized:

• ERA_A: 2017-01 through 2024-07 — original catalog structure
• ERA_B: 2024-08 through 2026-03 — updated catalog structure

The active parser (parse_catalog_v11.py) handles both eras. A full archive run produces 0 skipped editions and 0 body-parse anomalies.

The 2026-03 edition serves as the trusted reference baseline. It was validated deeply after an initial scrape returned incomplete results (696 AP codes vs. the correct 838). The discrepancy was traced, corrected, and verified.

14 structurally critical editions — breakpoints where the parser or catalog structure changed — were individually validated. All 14 passed clean.

WGU Atlas distinguishes between observed facts and inferred relationships:

• Observed: directly present in the catalog archive — course code, title, program membership, edition dates
• Inferred: derived from patterns across editions — event types, event interpretations, stability classifications

Interpretive content (event interpretations, title variant classifications) is labeled with confidence levels: high, moderate, or tentative.

167 course codes show title variation across editions. These have been manually classified into categories:

• Extraction noise (145 codes, 87%) — PDF line-wrap, Unicode variants, catalog oscillations. Not real renames.
• Punctuation only (16) — hyphen, comma, em-dash changes.
• Wording refinement (3) — typo fix or minor synonym swap.
• Substantive change (2) — genuine semantic renames.
• Formatting only (1) — space insertion.

The overwhelming majority of apparent title variation is extraction artifact, not editorial intent.

• Catalog date ≠ implementation dateThe catalog reflects publication timing, not guaranteed student rollout. A course appearing in a March catalog may have been deployed to students earlier or later.
• Catalog presence ≠ lived experienceOfficial structure does not perfectly capture actual student pathways or the subjective experience of a course.
• Code change ≠ substantive changeCourse code changes may reflect renumbering, administrative reorganization, or cleanup — not necessarily changes to course content.
• Reddit is supplementary contextStudent discussion data (planned for v1.1) is useful context, not institutional truth. WGU Atlas keeps official catalog facts and discussion signals in clearly separate sections.
• One-off courses require cautionCourses with only 1–2 catalog appearances are flagged (ghost_flag, single_appearance_flag). These may represent data anomalies or genuinely short-lived entries.

WGU Atlas enforces a strict separation between three information types:

• Official catalog facts — from the WGU public catalog archive
• Discussion signals — from Reddit and community spaces (v1.1)
• LLM-generated summaries — clearly labeled with date, source count, and disclaimer (v1.1)

These three types are never mixed in the same field or presented as equivalent.