815 messy customer records cleaned to 725 usable rows in Python

This is a self-directed demo, not client work. I wrote a Python pipeline with pandas that took a messy customer list of 815 rows and returned 725 clean records. It merged 54 duplicate rows and held 36 rows for review, each with a stated reason. No value was guessed.

Customer lists collect duplicates and broken values over time. A list in that state cannot go straight into a CRM or an email tool. Someone has to decide which rows are the same customer and which values can be trusted.

A Python script using pandas. It reads the messy CSV file, merges duplicate customer records and sets aside every row it cannot fix with confidence. Each held row gets a written reason, so a person can decide quickly.

Every input row is accounted for: 725 clean + 54 merged + 36 held = 815.

A row the script could not fix was held back with a reason, not filled with a guess.

815 rows in, 725 clean records out.

54 duplicate rows merged into the records they matched.

36 rows held for review, each with a stated reason.

Send me your messy CSV or Excel file. I merge duplicate records on a key you choose, fix inconsistent dates, casing and text, handle missing or malformed values and clean up the columns. You get the cleaned file, a short summary of what changed and a separate list of any records that need your decision. I never guess a value to fill a gap. The package covers up to about 5,000 rows. That is my "I will clean and organize your CSV or Excel data" service, from $30 on Contra.