Spreadsheet Data Cleaner & Deduplicator
Fix messy Excel and CSV files without touching the original. Profiles your data first, then standardizes phone numbers, dates (including ROC years and Excel serials), amounts like NT$12,000 or 1.2萬, emails, full-width text and spacing, splits mixed columns, finds exact and fuzzy duplicates, and merges files with different headers. Every change is logged; unclear values are listed for review.
#Productividad#Análisis#Comercio
Valoración
Se necesitan más valoraciones
Vendidos
0
Cómo se usa
Descargar
Clean messy spreadsheets without breaking them
Customer lists, order exports and survey results are rarely clean: phone numbers in five formats, dates in ROC years, amounts written as "1.2萬", the same customer entered three times. This Agent profiles your file, agrees a cleaning plan with you, and writes a new clean file with every change logged. Your original is never modified.
What it fixes
| Problem | Example → Result |
|---|---|
| Phone formats | 0912345678, +886 912-345-678, 912345678 → 0912-345-678 (keeps #ext) |
| Dates | 113/05/01 (ROC), 2024年5月1日, Excel serial 45413 → 2024-05-01 |
| Amounts | NT$12,000, 1.2萬, (500), 15% → 12000, 12000, -500, 0.15 |
| Text | full-width ＡＢＣ１２３, extra spaces, spelling variants (台北 / 臺北 / 台北市) |
| Columns | split "王小明 0912345678" into name + phone, merge columns, rename, drop |
| Duplicates | exact or fuzzy match on the columns you choose; keep the most complete row, or just flag |
| Several files | merge Jan/Feb/Mar exports even when headers differ slightly |
How it works
- Profile: a report of missing values, duplicates, inconsistent formats and suggested fixes.
- Plan: you confirm what to change (day-first or month-first dates, which columns identify a duplicate, drop or flag).
- Clean:
cleaned.xlsxwith three sheets (cleaned data, every change with before/after, values that need a human look) plus a step-by-step log.
Safe by design
- Leading zeros are kept (all cells written as text).
- Values the rules can't understand are left unchanged and listed, never guessed.
- Deleting rows always needs your confirmation.
Example
"Clean customers.xlsx: standardize phones and dates, and remove duplicate customers by name and phone."
Requirements
Python 3.8+ with pandas and openpyxl. Runs locally; your data stays on your machine.
Languages
Replies in Traditional Chinese, Simplified Chinese, English, Japanese, Korean, Vietnamese and Indonesian.
