Within-organization surname-rarity weight, per record

For each record, estimates how much information "the surname agrees" carries inside that record's organization. If every distinct person in an organization shares one surname (a family board) the factor approaches 0; a unique surname gives 1. Distinct people are approximated by distinct full-name strings so that the same person appearing in several years is not counted repeatedly.

Usage

surname_weight(df, org_id = "ein", name = "name", last_name = "last_name")

Arguments

df

Working data frame.

org_id, name, last_name

Column names for the organization id, the full name, and the surname.

Value

Numeric vector in [0, 1], one value per row of df.