Today felt like a shift from just learning concepts to actually building something practical.

Instead of focusing on small exercises, I worked on a real problem I’ve encountered many times in my experience as a WordPress and Laravel developer—messy WooCommerce product CSV files.

If you’ve ever imported products into WooCommerce, you probably know the pain:

• missing SKUs
• duplicate entries
• inconsistent formatting
• invalid stock statuses
• missing prices

These issues don’t just cause minor errors—they can break imports, mess up inventory, and create real problems in production.

So instead of fixing data manually inside WooCommerce, I wanted to solve the problem before the data even gets imported.

The Idea

I decided to build a simple Python tool that follows this flow:

CSV → Clean → Validate → Separate → Export → Report

The goal wasn’t to memorize Python syntax.
The goal was to solve a real problem using Python and AI as a coding partner.

How I Built It (Step-by-Step)

Instead of building everything at once, I approached this in small steps.

Step 1 — Reading the CSV File

I started by loading the CSV file using Pandas:

import pandas as pd

df = pd.read_csv("data/sample_products.csv")
print(df.head())

This helped me understand the structure of the data before doing anything else.

👉 At this point, I could already see the columns:

• name
• sku
• regular_price
• stock_status
• description
• categories

Step 2 — Cleaning the Data

Next, I cleaned the data to make it consistent:

df["name"] = df["name"].astype(str).str.strip()
df["sku"] = df["sku"].astype(str).str.strip().str.upper()
df["stock_status"] = df["stock_status"].astype(str).str.strip().str.lower()

This step fixes things like:

” tb-001 ” → “TB-001”
“In Stock” → “instock”

At this stage, I wasn’t rejecting anything yet—I was just fixing what could be fixed.

Step 3 — Validating the Data

This is where things got interesting.

I defined what “bad data” means:

valid_stock_status = ["instock", "outofstock", "onbackorder"]

df["missing_name"] = df["name"] == ""
df["missing_sku"] = df["sku"] == ""
df["missing_price"] = df["regular_price"].isna()
df["invalid_stock"] = ~df["stock_status"].isin(valid_stock_status)
df["duplicate_sku"] = df.duplicated(subset=["sku"], keep=False)

Instead of asking:

“Is this data correct?”

I started asking:

“Can my system trust this data?”

If not, it gets rejected.

Step 4 — Separating Clean and Invalid Data

I split the dataset into two:

valid_df = df[~invalid_mask]
invalid_df = df[invalid_mask]

• clean_products.csv → valid rows
• rejected_products.csv → invalid rows

This makes it easy to continue working only with clean data.

Step 5 — Adding Rejection Reasons

Instead of guessing what went wrong, I added a reason:

Missing SKU; Invalid stock status
Duplicate SKU
Missing price

This turns the script into something usable, not just technical.

Step 6 — Exporting Results

Finally, I exported the results:

valid_df.to_csv("output/clean_products.csv", index=False)
invalid_df.to_csv("output/rejected_products.csv", index=False)

Now I have:

• a clean dataset ready for import
• a rejected dataset for fixing

Step 7 — Generating a Summary Report

I also generated a simple report:

Total rows: 100
Valid rows: 72
Invalid rows: 28
Missing price: 8
Duplicate SKU: 10

This gives a quick overview of the data quality.

What I Learned

This project changed how I approach learning Python.

Instead of memorizing syntax, I focused on:

• defining a real problem
• using AI to generate code
• reviewing and refining the logic

As a developer, I realized:

The real value is not in writing code from memory—it’s in understanding the system and the data.

Real-World Applications

This type of tool is useful for:

• WooCommerce product imports
• ERP integrations
• API data validation pipelines
• preparing datasets for AI systems

It’s a simple project, but it solves a real problem I’ve encountered many times.

What’s Next

In Part 2, I’ll take this further by:

• turning this script into a simple web app
• allowing users to upload a CSV file
• automatically returning cleaned results

This is just the beginning—but it’s a solid step toward building real tools using Python and AI.