Rare Value Suppression

← Back
Current dataset: (none)
Scope
Choose whether to scan all columns or only the columns you pick below. Rare values are detected per column.
Threshold
  • By minimum count: treat values appearing fewer than m times as rare.
  • By minimum proportion: treat values with frequency below p% as rare.
Generalization Settings
Customize how rare values are generalized by type. Numeric options support fixed-width binning, rounding, or quantile binning (per-column based on observed distribution). Date options control time granularity. String options control how characters are masked or truncated.
Numbers
Example: width 10 → [0,10), [10,20)…
Per-column bin edges computed from data.
Dates
Non-ISO dates are parsed with browser Date; ambiguous formats may be skipped.
Strings
Rarity Preview

This preview estimates how many rows/values will be affected given your current scope and thresholds. It updates automatically when you change settings.

Action
Choose how to handle rare values: mask cells with *, drop rows entirely, or generalize values (use settings above).
Download Processed CSV