Lotte Duty Free Ranking Scraper — Korea Duty-Free Top 100

Scrape Lotte Duty Free (롯데면세점) — one of Korea's two dominant duty-free retailers — and get its Top-100 best-seller ranking, overall and for 14 categories, as clean JSON/CSV: rank, brand, product, USD list price, discount rate, discounted price, the KRW figure the board itself shows, rating and review count.

Built on the same ranking endpoint the kor.lottedfs.com storefront calls, so rows match what a shopper sees.

Why duty-free data? Duty-free boards are where Korean beauty, liquor and luxury demand from international shoppers shows first — and prices swing with promotions daily. Track the Top 100 every day and you see both the demand ranking and the real discounted price curve.

What you get

One row per ranked product:

Field	Meaning
ranking_category, rank	Which board, and the product's position (1–100)
name, brand	Product and brand
price_usd	List price
sale_price_usd, discount_rate	Discounted price when Lotte shows it (see below)
discount_hidden	true when Lotte hides the discount behind a login for this product
price_krw	The KRW figure the board itself renders
rating, review_count	Review score (5-point scale) and count
url, image_url, scraped_at	Product link, image, KST timestamp

The discount_hidden flag — read this

Lotte hides the discounted price of many products from logged-out visitors and shows "로그인 후 할인율 확인" instead. On a typical day that's roughly a third of the board (measured live: 505 of 1,500 rows). Those rows keep their real rank, list price and KRW figure, and carry discount_hidden: true with sale_price_usd/discount_rate null. That is Lotte's choice, not a scraping gap — the flag makes it explicit instead of papering over it.

What is deliberately NOT here

The final amount on Lotte's checkout (주문서) page also applies the shopper's own member grade, coupons and L.POINT balance — per-account values behind a login; no fixed "final price" exists on the site. This actor collects public data only: no login, no cart, no member or passport data.

Categories

전체(all) · 주류(liquor) · 스킨케어(skincare) · 메이크업(makeup) · 향수/바디/헤어(fragrance) · 뷰티 디바이스(beautydevice) · 가방/지갑(bags) · 시계/주얼리(watches) · 아이웨어/잡화(eyewear) · 패션/슈즈(fashion) · 스포츠/레져(sports) · 건강/가공식품(food) · 디지털/가전(digital) · 리빙/헬스케어(living) · 키즈/베이비(kids)

Leave categories empty to scrape all 15 boards (1,500 rows).

Input example

{
  "categories": ["전체", "스킨케어"],
  "maxItemsPerCategory": 100
}

Output example (real row)

{
  "source": "lottedfs",
  "ranking_category": "전체",
  "rank": 1,
  "name": "LG PraL SP 써마샷 얼티밋",
  "brand": "LG프라엘",
  "price_usd": 364,
  "sale_price_usd": 287.74,
  "discount_rate": 21,
  "discount_hidden": false,
  "price_krw": 406002,
  "rating": 4.5,
  "review_count": 51,
  "url": "https://kor.lottedfs.com/kr/product/20001034326"
}

Honest notes

• If a requested board returns nothing, the run warns loudly and records it in an EMPTY_CATEGORIES key-value record instead of exiting quietly.
• The run log summarises how many rows had their discount hidden, so a drop in coverage is visible immediately.
• No proxy needed for normal use; the board is served to ordinary IPs.