Name variants

• Names are typed by hand, and there is no single correct way to shorten a company name.
• Systems disagree on order: one stores family name first, another given name first.
• Legal suffixes, titles, and middle initials are included by some people and dropped by others.

• Totals per customer are split across the variants, so the largest customer can look like three small ones.
• Joins to another table match only the spelling that happens to agree.
• People receive duplicate mail, or mail addressed to a name they do not use.

• Sort the column alphabetically and scan for near neighbors.
• List distinct values after lowercasing and removing punctuation. The count drops sharply when variants exist.
• Group by a stable field such as email or tax ID and look for groups with more than one name.

• Separate formatting from identity. Case, spacing, and punctuation can be standardized safely.
• Decide a canonical form for suffixes and abbreviations and apply it consistently.
• Treat merging two differently named records as a decision that needs evidence beyond the name itself.

• Near-duplicate recordsTwo rows describe the same real thing but differ enough that an exact comparison misses them.
• Inconsistent category labelsA column meant to hold a few fixed values holds many spellings of them.
• Case example: Non-profitsA donor list that counted the same people twice
• Case example: Finance teamsA vendor list that paid the same invoice twice

TermsCanonical value, Entity resolution, Fuzzy matching, Standardization