You will get Clean and classify your eCommerce product CSV with QA flags
Project details
I will clean, normalize, and classify your eCommerce product CSV or spreadsheet into your fixed category list. This is useful when product titles are messy, duplicated, incomplete, or already mis-tagged. I use Python/pandas, rule-based checks, and AI-assisted review only where it helps. The output includes clean category assignments, normalized fields, and QA flags for rows that should be reviewed instead of guessed.
Machine Learning Tools
ChatGPT, Microsoft Excel, NumPy, pandas, Python, Scrapy, SQL, Tesseract OCRWhat's included
| Service Tiers |
Starter
$50
|
Standard
$180
|
Advanced
$450
|
|---|---|---|---|
| Delivery Time | 2 days | 4 days | 7 days |
Number of Revisions | 1 | 2 | 3 |
Number of Model Variations | 1 | 1 | 2 |
Number of Scenarios | 1 | 2 | 3 |
Number of Graphs/Charts | 0 | 1 | 2 |
Model Validation/Testing | |||
Model Documentation | |||
Data Source Connectivity | - | ||
Source Code | - |
Frequently asked questions
About Laiqing
eCommerce Product Data Cleanup | CSV & Python Automation
南昌市, China - 10:51 am local time
My work includes product categorization, CSV cleanup, Excel data cleaning, product listing cleanup, duplicate detection, title normalization, attribute cleanup, product image/file naming, and AI-assisted classification with human-review flags.
My main case study is a production product classification workflow built for Amazon return items. It has been used in real warehouse operations for almost four months and has processed 10,000+ real product records.
This is not a simple offline batch script. In daily warehouse use, operators scan product barcodes and the system classifies each item in real time using cached market data, rule-based logic, category mapping, and an LLM fallback. It flags uncertain or boundary cases for human review, records each decision, and learns from manual corrections.
Documented snapshot test:
- 1,178 consecutive real product entries
- 97.7% field acceptance rate
- 0 critical wrong submissions in that test
- 36 main categories and 156 subcategories
- Built with Python, APIs, caching, rules, LLM arbitration, and QA checks
What I can help with:
- Product categorization and taxonomy mapping
- CSV and spreadsheet cleanup
- Shopify, WooCommerce, Amazon, and eCommerce product import preparation
- Product listing cleanup and title normalization
- Attribute extraction and data normalization
- Duplicate detection
- Product image/file naming and basic Photoshop image cleanup
- Human-review workflows for uncertain rows
I can start with a small sample so you can validate quality before processing the full dataset.
Steps for completing your project
After purchasing the project, send requirements so Laiqing can start the project.
Delivery time starts when Laiqing receives requirements from you.
Laiqing works on your project following the steps below.
Revisions may occur after the delivery date.
Clean, classify, and flag uncertain rows
I review your schema and taxonomy, process the product file, assign categories from your fixed list, mark low-confidence rows for review, and deliver a clean CSV with notes.