README RESEARCHER INTELLIGENCE v1 0
# HE Open Data Ranking 2027 — Researcher-Level Institutional Intelligence v1.0
## Purpose
This module adds a **non-scoring dashboard layer** of up to 100 leading research contributors for every HE institution. It does **not** modify the frozen 24-KPI / 10,000-point university ranking methodology.
The primary unit remains the **university**. HE does not create a world, country, continental, or subject ranking of individual researchers in this module.
## Dashboard label
Recommended public label:
**Leading Research Contributors**
Recommended methodological subtitle:
**Researcher-Level Institutional Intelligence**
## HE Institutional Research Contribution Score (HE-IRCS)
Researchers are ordered inside each university profile using HE-IRCS (0–100):
- 40% — institution-affiliated citation impact percentile
- 25% — institution-affiliated h-index percentile
- 15% — institution-affiliated i10-index percentile
- 10% — institution-affiliated publication count percentile
- 10% — recent institutional citation impact percentile (2021–2025)
The affiliated metrics are calculated only from works where the **target author and target institution occur in the same OpenAlex authorship record**.
Lifetime OpenAlex citations, h-index, i10-index and works count are retained only as contextual fields; they are not the primary HE-IRCS inputs.
## Researcher association rule
Candidate researchers are identified using OpenAlex `last_known_institutions`. This is a bibliographic association signal and **does not necessarily prove current employment**.
HE should therefore avoid public labels such as "employees", "faculty members", or "staff" unless separately verified.
Recommended note:
> Researchers are associated with the institution using OpenAlex's most recently observed institutional affiliation. Affiliation metadata and author disambiguation may change as OpenAlex is updated.
## Candidate preselection
The module first preselects current-associated researchers by lifetime OpenAlex citation count, then calculates institution-specific HE-IRCS.
- Default module candidate limit: 500 per institution.
- API pilot BAT intentionally uses 150 for speed/cost.
- Production snapshot BAT uses 500.
- Increase `--candidate-limit` if greater recall is required.
A candidate cap is a computational prefilter, not an HE scientist rank.
## Eligibility
Default minimum: 3 exact institution-affiliated works.
If fewer than 100 researchers satisfy the rule, the dashboard should display the available number rather than artificially filling 100 positions.
## Production recommendation
For thousands of universities, use the **OpenAlex local JSONL snapshot** mode. API mode is suitable for pilots and small batches but is request-intensive.
### Required core input
`HE_Open_Data_Ranking_2027_Collection_v0_10_6.sqlite`
This is the SQLite database produced by the existing v0.10.6 collector.
### Production command
```bash
python he_researcher_intelligence_v1_0.py --mode snapshot-jsonl --snapshot-root "PATH_TO_OPENALEX_SNAPSHOT" --candidate-limit 500 --top-n 100
```
### Pilot API command
```bash
python he_researcher_intelligence_v1_0.py --mode api --candidate-limit 150 --top-n 100
```
The API mode uses batched author filters (up to 100 IDs per work-query batch) and verifies author + institution within the same authorship.
## Outputs
- `HE_Open_Data_Ranking_2027_Researcher_Intelligence_v1.sqlite`
- `HE_Open_Data_Ranking_2027_Researcher_Intelligence_v1.xlsx`
- `HE_Open_Data_Ranking_2027_Top100_Research_Contributors_v1.csv`
- `HE_Open_Data_Ranking_2027_Top100_Research_Contributors_v1.json`
The JSON output is intended for the public/private dashboard or API layer.
## Important governance rule
HE-IRCS is **not** a 25th KPI. It does not change:
- the official 10,000-point university score;
- global/country/continent rank;
- the five pillar scores;
- any of the 24 frozen KPIs.
This separation is deliberate so the new researcher-intelligence layer does not alter Methodology v1.0 after freeze.
## Difference from scientist-ranking products
HE uses researcher-level information only to explain the internal structure of a university's research performance. It does not create an independent global scientist league table.
The intended conceptual sequence is:
**University → institutional research ecosystem → leading research contributors**
not:
**Researcher → global scientist rank → aggregated university rank**.
## Purpose
This module adds a **non-scoring dashboard layer** of up to 100 leading research contributors for every HE institution. It does **not** modify the frozen 24-KPI / 10,000-point university ranking methodology.
The primary unit remains the **university**. HE does not create a world, country, continental, or subject ranking of individual researchers in this module.
## Dashboard label
Recommended public label:
**Leading Research Contributors**
Recommended methodological subtitle:
**Researcher-Level Institutional Intelligence**
## HE Institutional Research Contribution Score (HE-IRCS)
Researchers are ordered inside each university profile using HE-IRCS (0–100):
- 40% — institution-affiliated citation impact percentile
- 25% — institution-affiliated h-index percentile
- 15% — institution-affiliated i10-index percentile
- 10% — institution-affiliated publication count percentile
- 10% — recent institutional citation impact percentile (2021–2025)
The affiliated metrics are calculated only from works where the **target author and target institution occur in the same OpenAlex authorship record**.
Lifetime OpenAlex citations, h-index, i10-index and works count are retained only as contextual fields; they are not the primary HE-IRCS inputs.
## Researcher association rule
Candidate researchers are identified using OpenAlex `last_known_institutions`. This is a bibliographic association signal and **does not necessarily prove current employment**.
HE should therefore avoid public labels such as "employees", "faculty members", or "staff" unless separately verified.
Recommended note:
> Researchers are associated with the institution using OpenAlex's most recently observed institutional affiliation. Affiliation metadata and author disambiguation may change as OpenAlex is updated.
## Candidate preselection
The module first preselects current-associated researchers by lifetime OpenAlex citation count, then calculates institution-specific HE-IRCS.
- Default module candidate limit: 500 per institution.
- API pilot BAT intentionally uses 150 for speed/cost.
- Production snapshot BAT uses 500.
- Increase `--candidate-limit` if greater recall is required.
A candidate cap is a computational prefilter, not an HE scientist rank.
## Eligibility
Default minimum: 3 exact institution-affiliated works.
If fewer than 100 researchers satisfy the rule, the dashboard should display the available number rather than artificially filling 100 positions.
## Production recommendation
For thousands of universities, use the **OpenAlex local JSONL snapshot** mode. API mode is suitable for pilots and small batches but is request-intensive.
### Required core input
`HE_Open_Data_Ranking_2027_Collection_v0_10_6.sqlite`
This is the SQLite database produced by the existing v0.10.6 collector.
### Production command
```bash
python he_researcher_intelligence_v1_0.py --mode snapshot-jsonl --snapshot-root "PATH_TO_OPENALEX_SNAPSHOT" --candidate-limit 500 --top-n 100
```
### Pilot API command
```bash
python he_researcher_intelligence_v1_0.py --mode api --candidate-limit 150 --top-n 100
```
The API mode uses batched author filters (up to 100 IDs per work-query batch) and verifies author + institution within the same authorship.
## Outputs
- `HE_Open_Data_Ranking_2027_Researcher_Intelligence_v1.sqlite`
- `HE_Open_Data_Ranking_2027_Researcher_Intelligence_v1.xlsx`
- `HE_Open_Data_Ranking_2027_Top100_Research_Contributors_v1.csv`
- `HE_Open_Data_Ranking_2027_Top100_Research_Contributors_v1.json`
The JSON output is intended for the public/private dashboard or API layer.
## Important governance rule
HE-IRCS is **not** a 25th KPI. It does not change:
- the official 10,000-point university score;
- global/country/continent rank;
- the five pillar scores;
- any of the 24 frozen KPIs.
This separation is deliberate so the new researcher-intelligence layer does not alter Methodology v1.0 after freeze.
## Difference from scientist-ranking products
HE uses researcher-level information only to explain the internal structure of a university's research performance. It does not create an independent global scientist league table.
The intended conceptual sequence is:
**University → institutional research ecosystem → leading research contributors**
not:
**Researcher → global scientist rank → aggregated university rank**.