Open Dataset Profiler
Drop in any tabular CSV - health records, hiring data, loan applications - and get an instant demographic representation audit: under-represented or missing subgroups, skewed age and sex distributions, geographic under-sampling, and intersectional gaps.
Your file never leaves your browser. All analysis runs locally on this page - nothing is uploaded.
Drop a CSV, TSV, JSON, or Excel file here, or browse
Comma- or tab-separated values with a header row, JSON
(records or split orientation), or an .xlsx workbook (first sheet). (Parquet files: use
faircode profile data.parquet from the CLI.) Proxy-hint detection: use
faircode profile --proxy-hints from the CLI.
Flags
Column mapping auto-detection missed a column? map it here
Detection keys off column names, so an oddly-named column
(gndr, patient_region_code) can be missed or mistyped. Override any
column below and the audit re-runs instantly - nothing leaves your browser.
Detected demographic dimensions
Intersectional gaps
Compare two datasets - representation drift
Profile a baseline against a later snapshot - training vs. production, last quarter vs. this one - to see how each demographic dimension's representation shifted. Drift is measured with the Population Stability Index (PSI). Both files stay in your browser.
Column mapping auto-detection missed a column? map it here
Applies to both datasets A and B - override any column below and the comparison re-runs instantly on both sides, nothing leaves your browser.