card

README.md

@@ -0,0 +1,50 @@
+---
+license: cc0-1.0
+language: [en]
+task_categories: [text-retrieval, tabular-classification]
+tags: [legal, law, judges, judiciary, us-courts, public-domain, judicial-analytics]
+pretty_name: US Judges (CourtListener people database)
+configs:
+- config_name: judges
+  data_files: data/judges/judges.jsonl.gz
+  default: true
+- config_name: positions
+  data_files: data/positions/positions.jsonl.gz
+- config_name: educations
+  data_files: data/educations/educations.jsonl.gz
+- config_name: schools
+  data_files: data/schools/schools.jsonl.gz
+- config_name: political_affiliations
+  data_files: data/political_affiliations/political_affiliations.jsonl.gz
+- config_name: retention_events
+  data_files: data/retention_events/retention_events.jsonl.gz
+---
+
+# US Judges — the CourtListener people database
+
+Structured records of U.S. judges: who they are, every judicial position they have held, where they
+studied, and their political affiliations. From the Free Law Project / CourtListener bulk export of
+**2026-06-30**, public domain.
+
+| config | rows |
+|---|---:|
+| `judges` | 16,191 |
+| `positions` | 51,291 |
+| `educations` | 12,777 |
+| `schools` | 6,011 |
+| `political_affiliations` | 8,486 |
+| `retention_events` | 0 |
+
+Join `positions` to `judges` on the person id to trace a judge's career across courts. Coverage is
+strongest for federal judges and state appellate judges; many state trial judges are not in this
+database, and for those the opinions themselves (see `docketx/us-caselaw`) name the presiding judge.
+
+## Load it
+```python
+from datasets import load_dataset
+judges = load_dataset("docketx/us-judges", "judges")
+positions = load_dataset("docketx/us-judges", "positions")
+```
+
+Part of the [DocketX](https://docketx.ai) open legal corpus, alongside `docketx/us-caselaw`
+(the national opinion corpus) and `docketx/oral-arguments-us` (argument transcripts).