20,000 SKU Mapping for a UAE Electronics Retailer: A Complete Technical Tutorial

Introduction

SKU mapping is one of the hardest problems in retail analytics. Electronics especially — because the same product appears in dozens of formats:

• different titles
• different model numbers
• different sellers
• different RAM + storage combos
• different colors
• different accessories included
• different warranty types
• different bundle packs

A single smartphone like Samsung Galaxy A34 5G can appear in 35–50 variations across:

• Noon UAE
• Amazon.ae
• Carrefour UAE
• Sharaf DG
• Emax
• Jumbo Electronics

If you're mapping 20,000 SKUs, manual processing is impossible.

This tutorial shows you exactly how to:

• scrape product listings
• extract SKU attributes
• clean & normalize raw data
• extract model codes
• link variants
• build a product-family mapping
• output the final dataset

This is the same workflow Actowiz Solutions uses for large-scale retail analytics projects for GCC electronics retailers.

Step 1: Install Your Required Tools

Install all external libraries:

pip install selenium
pip install pandas
pip install fuzzywuzzy
pip install python-Levenshtein
pip install requests
pip install beautifulsoup4

You will use:

• Selenium → website loading
• Requests/BS4 → HTML parsing
• Pandas → data structuring
• FuzzyWuzzy → similarity matching
• Custom Python functions → model code extraction

Step 2: Scrape Electronics Products From a Retailer (Example: Amazon.ae)

Basic imports:

from selenium import webdriver
from selenium.webdriver.common.by import By
from selenium.webdriver.common.keys import Keys
import pandas as pd
from time import sleep
import re
from fuzzywuzzy import fuzz
import json

Open product category:

browser = webdriver.Chrome()
browser.get("https://www.amazon.ae/s?k=smartphones")
sleep(3)

Scroll for pagination:

for _ in range(8):
    browser.find_element(By.TAG_NAME, "body").send_keys(Keys.END)
    sleep(2)

Step 3: Extract Product Title, Price, Rating, and Link

items = browser.find_elements(By.XPATH, '//div[@data-component-type="s-search-result"]')

records = []

for item in items:
    try:
        title = item.find_element(By.TAG_NAME, "h2").text
    except:
        title = ""

    try:
        price = item.find_element(By.CLASS_NAME, "a-price-whole").text
    except:
        price = ""

    try:
        link = item.find_element(By.TAG_NAME, "a").get_attribute("href")
    except:
        link = ""

    records.append({
        "title": title,
        "price": price,
        "url": link
    })

Export the first batch:

df = pd.DataFrame(records)
df.head()

This becomes your raw SKU dataset.

Step 4: Extract Model Numbers Using Regex

Electronics SKUs always contain model codes like:

• SM-A546E
• iPhone 13 A2633
• MGN93HN/A
• CPH2551
• Redmi 12C

Use regex:

def extract_model(title):
    pattern = r"([A-Za-z0-9-]{4,})"
    matches = re.findall(pattern, title)
    return matches

df["model_codes"] = df["title"].apply(extract_model)

Example result:

This is the most important step in SKU mapping.

Step 5: Normalize Titles Across Retailers

Remove noise words:

def clean_title(t):
    t = t.lower()
    remove = ["official", "2025 version", "brand new", "uae version", "global version"]
    for r in remove:
        t = t.replace(r, "")
    return t.strip()

df["clean_title"] = df["title"].apply(clean_title)

This ensures consistency across platforms.

Step 6: Fuzzy Match Titles to Find Similar SKUs

For SKU mapping, we need to identify when:

• Noon title = Amazon title
• Amazon title = Carrefour title
• Jumbo title = Sharaf DG title

FuzzyWuzzy helps:

def is_match(t1, t2, threshold=80):
    score = fuzz.token_set_ratio(t1, t2)
    return score >= threshold

Example:

is_match("samsung galaxy a54 5g", "galaxy a54 samsung phone 5g")
# Output: True

Step 7: Build SKU Groups Based on Model Code + Fuzzy Match

Here is a simple grouping logic:

sku_groups = {}

for i, row in df.iterrows():
    model = tuple(row["model_codes"])

    if model not in sku_groups:
        sku_groups[model] = []

    sku_groups[model].append({
        "title": row["title"],
        "price": row["price"],
        "url": row["url"],
        "clean_title": row["clean_title"]
    })

This ensures all SKUs of the same model gather into one family.

Step 8: Generate the Final SKU Mapping Output

with open("sku_mapping_output.json", "w") as f:
    json.dump(sku_groups, f, indent=4)

Final structure looks like:

{
  "('SM-A546E',)": [
    {
      "title": "Samsung Galaxy A54 5G 8GB 256GB SM-A546E",
      "price": "1199",
      "url": "https://www.amazon.ae/..."
    },
    {
      "title": "Samsung A54 5G Dual SIM 256GB",
      "price": "1150",
      "url": "https://www.noon.com/..."
    }
  ]
}

This is SKU mapping for one product family.

Repeat until you process 20,000 SKUs.

Step 9: Add Optional Enhancements

• Color extraction
• Using regex or image scraping.
• Storage classification
• Detect 128GB vs 256GB variations.
• RAM configuration detection
• 8GB / 12GB versions.
• Region mapping
• UAE version vs KSA version.
• Price analysis
• Median, max, min, volatility.
• Duplicate listings
• Remove exact duplicates.

These are crucial for retailers.

Step 10: Challenges You Will Face

This tutorial script works, but large-scale SKU mapping faces issues:

Retailer websites block scraping

Requires proxies, residential IPs, and anti-bot solutions.

Electronics model codes vary heavily

Requires smarter regex + NLP-based extraction.

20,000 SKUs require distributed crawling

Local machine will not handle this scale.

Fuzzy matching gets slow

Requires vector-based similarity using sentence transformers.

Data accuracy is critical

Slight errors can mis-map SKUs.

When Should You Use Actowiz Solutions Instead of DIY Scripts?

If your project involves:

• 10,000+ SKUs
• daily price/stock updates
• multi-retailer extraction
• multi-country catalog mapping
• integrating with BI systems
• variant-level insights
• historical trend analysis

…then manual coding is not enough.

Actowiz Solutions provides:

• Automated SKU mapping engine
• AI-based similarity matching
• Multi-marketplace extraction (Noon, Amazon, Sharaf DG)
• Full catalog normalization
• Daily/real-time price updates
• Enterprise-grade dashboards
• On-demand API delivery

Electronics, fashion, grocery — Actowiz maps SKUs across all industries.

Conclusion

Mapping 20,000 SKUs across multiple electronics retailers is a complex problem involving:

• scraping
• cleaning
• normalization
• pattern detection
• model extraction
• fuzzy matching
• grouping
• output formatting

The tutorial above gives you the complete technical pipeline to build your own SKU mapping solution.

But if you want an automated, scalable, production-ready system for UAE retail — Actowiz Solutions can deploy a fully managed SKU mapping framework tailored to your needs.

You can also reach us for all your mobile app scraping, data collection, web scraping , and instant data scraper service requirements!