Data Cleansing & Processing Suite

Ingest real client data. Run deterministic cleaning rules — phone formats, ZIP codes, email typos, fuzzy duplicate merging. No unnecessary AI inference. No black boxes. Every correction is a visible, auditable rule.

1. Select Production-Grade Dataset (150 Rows) or Ingest Custom CSV

Non-Profit Donor Operations & CASS Hygiene

Dataset Context: 150 donor records with malformed phone numbers (+1 E.164), dropped leading zeros on East Coast ZIP codes, email domain typos (@gmai.com), and phonetic household duplicates.

Hygiene Score Lift

68%──►98%

Static baseline ──► Dynamic clean state

Format Cells Repaired38Phone, date, zip, email, whitespace

Fuzzy Duplicates Pruned3Jaro-Winkler entity matches

Certified Clean Export

147

Pruned duplicate entities

Modular Cleansing Rule Engine & Script Selection

Whitespace TrimE.164 PhoneEmail Typo FixZIP 0-PadderISO Date / ExcelOCR / UOM Lock

Fuzzy Deduplication Similarity Threshold (τ): 85%Balanced (Recommended)Calibrates Jaro-Winkler entity distance to govern how aggressively duplicate records are merged.

50% (Aggressive)100% (Strict)

How Threshold (τ) Governs Record Merging & Clean Score:

85% (Balanced - Standard Enterprise Gate): Successfully identifies and merges standard business variations (e.g. "Acme Global Inc" vs "Acme-Global", or "Bob & Mary Smith" vs "Robert Smith") with near-zero false merges, delivering optimal data hygiene.

Tabular Inspection & Transformation Diff View

Green = Deterministically Fixed | Amber = Warning/Anomaly | Red Strikethrough = Duplicate Pruned

Status	Donor_ID	First_Name	Last_Name	Spouse_Name	Address_Line1	City	State	ZIP_Code	Phone	Email	Giving_Tier	Total_Lifetime_Donations	Last_Donation_Date	DAF_Vehicle

Showing 0 of 0 rowsColumns: 14