apify-ecommerce – skill Dữ liệu & Khoa học cho agent AI

E-commerce Data Extraction

Extract product data, prices, reviews, and seller information from any e-commerce platform using Apify's E-commerce Scraping Tool.

When to Use

• You need product, pricing, review, stock, or seller data from e-commerce sites.
• The task involves price monitoring, competitor product comparison, MAP enforcement, or review analysis.
• You need a guided workflow for extracting marketplace data and summarizing findings.

Prerequisites

• .env file with APIFY_TOKEN (at ~/.claude/.env)
• Node.js 20.6+ (for native --env-file support)

Workflow Selection

| User Need | Workflow | Best For |

|-----------|----------|----------|

| Track prices, compare products | Workflow 1: Products & Pricing | Price monitoring, MAP compliance, competitor analysis. Add AI summary for insights. |

| Analyze reviews (sentiment or quality) | Workflow 2: Reviews | Brand perception, customer sentiment, quality issues, defect patterns |

| Find sellers across stores | Workflow 3: Sellers | Unauthorized resellers, vendor discovery via Google Shopping |

Progress Tracking

Task Progress:
- [ ] Step 1: Select workflow and determine data source
- [ ] Step 2: Configure Actor input
- [ ] Step 3: Ask user preferences (format, filename)
- [ ] Step 4: Run the extraction script
- [ ] Step 5: Summarize results

---

Workflow 1: Products & Pricing

Use case: Extract product data, prices, and stock status. Track competitor prices, detect MAP violations, benchmark products, or research markets.

Best for: Pricing analysts, product managers, market researchers.

Input Options

| Input Type | Field | Description |

|------------|-------|-------------|

| Product URLs | detailsUrls | Direct URLs to product pages (use object format) |

| Category URLs | listingUrls | URLs to category/search result pages |

| Keyword Search | keyword + marketplaces | Search term across selected marketplaces |

Example - Product URLs

{
  "detailsUrls": [
    {"url": "https://www.amazon.com/dp/B09V3KXJPB"},
    {"url": "https://www.walmart.com/ip/123456789"}
  ],
  "additionalProperties": true
}

Example - Keyword Search

{
  "keyword": "Samsung Galaxy S24",
  "marketplaces": ["www.amazon.com", "www.walmart.com"],
  "additionalProperties": true,
  "maxProductResults": 50
}

Optional: AI Summary

Add these fields to get AI-generated insights:

| Field | Description |

|-------|-------------|

| fieldsToAnalyze | Data points to analyze: ["name", "offers", "brand", "description"] |

| customPrompt | Custom analysis instructions |

Example with AI summary:

{
  "keyword": "robot vacuum",
  "marketplaces": ["www.amazon.com"],
  "maxProductResults": 50,
  "additionalProperties": true,
  "fieldsToAnalyze": ["name", "offers", "brand"],
  "customPrompt": "Summarize price range and identify top brands"
}

Output Fields

• name - Product name
• url - Product URL
• offers.price - Current price
• offers.priceCurrency - Currency code (may vary by seller region)
• brand.slogan - Brand name (nested in object)
• image - Product image URL
• Additional seller/stock info when additionalProperties: true

> Note: Currency may vary in results even for US searches, as prices reflect different seller regions.

---

Workflow 2: Customer Reviews

Use case: Extract reviews for sentiment analysis, brand perception monitoring, or quality issue detection.

Best for: Brand managers, customer experience teams, QA teams, product managers.

Input Options

| Input Type | Field | Description |

|------------|-------|-------------|

| Product URLs | reviewListingUrls | Product pages to extract reviews from |

| Keyword Search | keywordReviews + marketplacesReviews | Search for product reviews by keyword |

Example - Extract Reviews from Product

{
  "reviewListingUrls": [
    {"url": "https://www.amazon.com/dp/B09V3KXJPB"}
  ],
  "sortReview": "Most recent",
  "additionalReviewProperties": true,
  "maxReviewResults": 500
}

Example - Keyword Search

{
  "keywordReviews": "wireless earbuds",
  "marketplacesReviews": ["www.amazon.com"],
  "sortReview": "Most recent",
  "additionalReviewProperties": true,
  "maxReviewResults": 200
}

Sort Options

• Most recent - Latest reviews first (recommended)
• Most relevant - Platform default relevance
• Most helpful - Highest voted reviews
• Highest rated - 5-star reviews first
• Lowest rated - 1-star reviews first

> Note: The sortReview: "Lowest rated" option may not work consistently across all marketplaces. For quality analysis, collect a large sample and filter by rating in post-processing.

Quality Analysis Tips

• Choose a bounded sampling plan; more reviews alone do not establish statistical significance or representative sentiment
• Look for recurring keywords: "broke", "defect", "quality", "returned"
• Filter results by rating if sorting doesn't work as expected
• Cross-reference with competitor products for benchmarking

---

Workflow 3: Seller Intelligence

Use case: Find sellers across stores, discover unauthorized resellers, evaluate vendor options.

Best for: Brand protection teams, procurement, supply chain managers.

> Note: This workflow uses Google Shopping to find sellers across stores. Direct seller profile URLs are not reliably supported.

Input Configuration

{
  "googleShoppingSearchKeyword": "Nike Air Max 90",
  "scrapeSellersFromGoogleShopping": true,
  "countryCode": "us",
  "maxGoogleShoppingSellersPerProduct": 20,
  "maxGoogleShoppingResults": 100
}

Options

| Field | Description |

|-------|-------------|

| googleShoppingSearchKeyword | Product name to search |

| scrapeSellersFromGoogleShopping | Set to true to extract sellers |

| scrapeProductsFromGoogleShopping | Set to true to also extract product details |

| countryCode | Target country (e.g., us, uk, de) |

| maxGoogleShoppingSellersPerProduct | Max sellers per product |

| maxGoogleShoppingResults | Total result limit |

---

Supported Marketplaces

Amazon (20+ regions)

www.amazon.com, www.amazon.co.uk, www.amazon.de, www.amazon.fr, www.amazon.it, www.amazon.es, www.amazon.ca, www.amazon.com.au, www.amazon.co.jp, www.amazon.in, www.amazon.com.br, www.amazon.com.mx, www.amazon.nl, www.amazon.pl, www.amazon.se, www.amazon.ae, www.amazon.sa, www.amazon.sg, www.amazon.com.tr, www.amazon.eg

Major US Retailers

www.walmart.com, www.costco.com, www.costco.ca, www.homedepot.com

European Retailers

allegro.pl, allegro.cz, allegro.sk, www.alza.cz, www.alza.sk, www.alza.de, www.alza.at, www.alza.hu, www.kaufland.de, www.kaufland.pl, www.kaufland.cz, www.kaufland.sk, www.kaufland.at, www.kaufland.fr, www.kaufland.it, www.cdiscount.com

IKEA (40+ country/language combinations)

Supports all major IKEA regional sites with multiple language options.

Google Shopping

Use for seller discovery across multiple stores.

---

Running the Extraction

Step 1: Set Skill Path

SKILL_PATH=~/.claude/skills/apify-ecommerce

Step 2: Run Script

Quick answer (display in chat):

node --env-file="$HOME/.claude/.env" "$SKILL_PATH/reference/scripts/run_actor.js" \
  --actor "apify/e-commerce-scraping-tool" \
  --input 'JSON_INPUT'

CSV export:

node --env-file="$HOME/.claude/.env" "$SKILL_PATH/reference/scripts/run_actor.js" \
  --actor "apify/e-commerce-scraping-tool" \
  --input 'JSON_INPUT' \
  --output YYYY-MM-DD_filename.csv \
  --format csv

JSON export:

node --env-file="$HOME/.claude/.env" "$SKILL_PATH/reference/scripts/run_actor.js" \
  --actor "apify/e-commerce-scraping-tool" \
  --input 'JSON_INPUT' \
  --output YYYY-MM-DD_filename.json \
  --format json

Step 3: Summarize Results

Report:

• Number of items extracted
• File location (if exported)
• Key insights based on workflow:
• Products: Price range, outliers, MAP violations
• Reviews: Average rating, sentiment trends, quality issues
• Sellers: Seller count, seller candidates found; authorization requires separate brand/contract evidence

---

Error Handling

| Error | Solution |

|-------|----------|

| APIFY_TOKEN not found | Ensure ~/.claude/.env contains APIFY_TOKEN=your_token |

| Actor not found | Verify Actor ID: apify/e-commerce-scraping-tool |

| Run FAILED | Check Apify console link in error output |

| Timeout | Reduce maxProductResults or increase --timeout |

| No results | Verify URLs are valid and accessible |

| Invalid marketplace | Check marketplace value matches supported list exactly |

Limitations

• Use this skill only when the task clearly matches the scope described above.
• Do not treat the output as a substitute for environment-specific validation, testing, or expert review.
• Stop a

… (nội dung đầy đủ có trong gói tải về)