1) Data Sources & Joins

University Staff Directories

• Scrape official Accounting/Finance staff pages across G8 universities.
• Collected fields: researcher name, job title, profile URL, field (Accounting/Finance).

Journal Quality Lists (ABDC/JQL)

• Load ABDC/JQL files (e.g., 2010/2016/2022).
• Primary join on ISSN (preferred).
• Fallback to fuzzy title matching when ISSN is missing/ambiguous.

Journal Impact Factor (Clarivate JIF)

• Where ISSN exists in JCR, attach JIF and 5-year JIF to the journal.
• Journals without JIF are excluded from JIF-based averages.

Database Linking (journal matching pass)

• Match Publications.journal_name → Journals using fuzzywuzzy process.extractOne with threshold 95.
• force=True: reset existing journal_id before re-matching.
• university="all" | "<name>": limit matching scope.

2) Standardisation Pipeline (Exactly How We Clean & Normalize)

2.1 Required fields & null handling

• Required: Title (index 0), Type (2), Researcher Name (5), Profile URL (6). Missing → row rejected.
• Empty → None for: Year (1), Journal Name (3), Article URL (4).

2.2 Type cleanup

• Trim trailing › from publication type strings where present.

2.3 Year coercion

• If Year is numeric, convert to integer.

2.4 Role blacklist (exclusions)

• If job title contains any of: Education-Focused, Education Focused, Education Focussed, Teaching-Focused, Teaching Focused, Teaching Focussed → set job title to "Exclude".
• Rationale: these positions are not comparable to research-active roles for ranking.

2.5 Title normalization (canonical roles)

Normalize raw titles to canonical forms. We search Job Title first, then Researcher Name if needed.

Raw / Variant	Canonical
Associate Lecturer	Associate Lecturer
Lecturer (A)	Associate Lecturer
Lecturer	Lecturer
Fellow	Fellow
Senior Lecturer	Senior Lecturer
Senior Fellow	Senior Fellow
Associate Professor / Associate Prof / AsPr	Associate Professor
Professor / Prof	Professor
Professorial Fellow	Professorial Fellow
Professor Emeritus / Emeritus Professor / Emeritus	Professor Emeritus

2.6 Name cleaning

• Strip prefixes from Researcher Name: Dr., Ms., Mr., Mrs., Prof., Associate Professor, Professor, Scientia Professor, Professor Scientia, Emeritus Professor, Professor Emeritus, Emeritus, and short codes EmPr, AsPr.
• Result example: “Prof. Jane Q. Smith” → “Jane Q. Smith”.

2.7 Academic level mapping (A–E)

Canonical Title	Level
Associate Lecturer	A
Lecturer	B
Fellow	B
Senior Lecturer	C
Senior Fellow	C
Associate Professor	D
Professor	E
Professorial Fellow	E
Professor Emeritus	E
Exclude	—

Note: If role is None or unrecognized, the level is set to None.

3) What Gets Written to the Database

• After standardization, the academic level is inserted into each row at index 9.
• Skip researchers whose job title equals "Exclude".
• Researcher uniqueness: (Name, Profile URL). If found, update job title / level / field when changed.
• Publication uniqueness: (Title, Researcher). New publications are inserted and linked.

4) Ranking Metrics (Exactly How They're Computed)

1. Total Publications: count of journal articles retrieved (after cleaning).
2. A*/A-Ranked Publications: count of publications whose journals map to ABDC A* or A.
3. Average JIF: mean of JIF values across only those articles with valid JIF.
   Avg JIF = (Σ JIFi) / N, where N = number of articles with JIF.
4. Average 5-Year JIF: same as above, using 5-year JIF.
5. Average Citation Percentile: mean of available OpenAlex citation percentiles for a researcher’s publications.

Notes: JIF and 5-year JIF are journal-level stats joined via ISSN → JCR. If no JIF exists, JIF-based averages are undefined (not zero).

5) Known Limitations (Transparency for Users)

• University data dependency: institutional directories may be incomplete/out-of-date.
• Name/affiliation matching noise: can produce over-/under-counting, especially for common names.
• Journal matching: fuzzy matching (threshold 95) can mis-map near-duplicates; ISSN is preferred but not always present.
• Coverage gaps: some journals lack ABDC ranks or JIF, leading to missing metrics.
• Metric bias: total publications favor volume; averages can be skewed by outliers; citation norms vary by field/year.

6) Worked Example (End-to-End)

Assume a researcher has 5 articles:

• Four with JIFs: 7.5, 5.0, 2.5, 1.2; ABDC: A*, A, A, B.
• One in a journal with no JIF and no ABDC entry.

Totals: total publications = 5; A*/A publications = 3.

Avg JIF: (7.5 + 5.0 + 2.5 + 1.2) / 4 = 4.05

Avg 5-year JIF: computed the same way, using 5-year values only.

Avg Citation Percentile: mean of all available publication percentiles.

7) Complete Lists (for Auditing)

Accepted canonical roles

• Associate Lecturer
• Lecturer
• Fellow
• Senior Lecturer
• Senior Fellow
• Associate Professor
• Professor
• Professorial Fellow
• Professor Emeritus

Raw variants recognized (mapped to above)

• Associate Lecturer, Lecturer (A), Lecturer, Fellow, Senior Lecturer, Senior Fellow
• Associate Professor, Associate Prof, AsPr
• Professor, Prof, Professorial Fellow
• Professor Emeritus, Emeritus Professor, Emeritus

Excluded keywords (blacklist)

• Education-Focused, Education Focused, Education Focussed
• Teaching-Focused, Teaching Focused, Teaching Focussed