About this data

Source

UGA Registrar Schedule of Classes — one machine-readable CSV (Spring 2026) plus the historical PDF archive — cross-referenced with the Open Georgia salary export.

Coverage: FY2015–FY2025 (33 historical PDF terms) plus a Spring 2026 CSV; salary FY2015–FY2025.

How it was collected

• The Spring 2026 CSV carries full instructor names, department, college, enrollment and cross-list, and links directly to salary.
• 33 historical Schedule-of-Classes PDFs (FY2015–2025) are parsed by detecting each page's column layout (positions drift across eras) and handling both 12- and 24-hour meeting times.
• PDFs give the instructor's LAST NAME ONLY, so linking is roster-assisted: last name + department against a roster built from the CSV and salary list; ambiguous names are left unlinked (precision over recall).
• A department-keyed back-propagation pass fills in instructors whose (surname, department) resolves uniquely elsewhere in the data.

Accuracy & instructor linking

• CSV term: about 81% of instructors linked. Historical PDF terms: about 57% of instructors linked.
• The gap is inherent to last-name-only PDFs — common surnames shared by multiple salaried people cannot be told apart without a first name.

Limitations

• Roughly 18% of PDF sections list no instructor at all, and many more carry an ambiguous last-name-only instructor that cannot be safely resolved.
• Only the CSV term carries a cross-list column, so co-scheduled sections are de-duplicated only there; PDF terms may slightly over-count co-scheduled classes.
• Department and college come directly from the source; contact hours are derived from meeting times.
• Resolving more historical instructors would require an external source with first names (e.g. an open-records extract).