Dataset

clinical-trials-trec-parsed

Public Records & Filings Download 1 source

This dataset is a parsed snapshot of ClinicalTrials.gov records, distributed in Parquet format on the Hugging Face Hub under an unspecified license. It contains roughly 52,000 trial entries drawn from public registry filings.

Open on Hugging Face

Data preview

Subset

Loading preview...

Dataset card

What this dataset contains

The release covers about 51,939 rows of structured clinical trial metadata, sitting in the 10K–100K size band typical of mid-scale text datasets. Each row corresponds to a single registered trial and carries descriptive text alongside eligibility and demographic fields. The publisher (2001jdev) does not document a refresh cadence, so the snapshot should be treated as a point-in-time extraction.

Structure and fields

Ten string-typed columns are exposed: trial_id, brief_title, brief_summary, inclusion_criteria, exclusion_criteria, phase, url, minimum_age, maximum_age, and gender. The phase field captures combined phases such as "1/2", while age bounds are stored as free-text strings like "14 Years" and "35 Years". trial_id values follow the NCT identifier convention used on ClinicalTrials.gov, and url points back to the canonical registry page for each study.

How analysts use it

Healthcare-focused analysts can profile trial pipelines by aggregating brief_title and phase to track how many studies target a given indication or mechanism. NLP pipelines can train on inclusion_criteria, exclusion_criteria, and brief_summary to classify trials or extract endpoints. Investors monitoring biotech catalysts may also filter by phase and gender or minimum_age to flag late-stage assets likely to read out.

Documentation & references 1