Datasetpublic

jeffreyszhou/ikea-us-products-2025

IKEA US Product Dataset (July 2025) This dataset is a structured snapshot of ~30,000 IKEA US products, scraped from the official IKEA US website in July 2025. It contains product metadata (titles, descriptions, categories, materials, care instructions, etc.) and associated product images. Contents products-us.jsonl — one JSON object per product with structured fields. images-us/ — the first "hero" image for each product, downloaded via image_downloader_first.py.… See the full description on the dataset page: https://huggingface.co/datasets/jeffreyszhou/ikea-us-products-2025.

sourceHugging Facemitupdated 1y agoView on Hugging Face

5likes118downloads

Dataset Card

IKEA US Product Dataset (July 2025)

This dataset is a structured snapshot of \~30,000 IKEA US products, scraped from the official IKEA US website in July 2025. It contains product metadata (titles, descriptions, categories, materials, care instructions, etc.) and associated product images.

—products-us.jsonl — one JSON object per product with structured fields.
—images-us/ — the first "hero" image for each product, downloaded via image_downloader_first.py. (Optional: use image_downloader_all.py if you want all images per product.)
—scraper/ — Python scripts used to build the dataset.
—README.md (this file).

Schema

Each line in products-us.jsonl is a JSON object with the following fields:

Field	Type	Description
`product_id`	`string`	8-digit IKEA product identifier.
`title`	`string`	Product name + size/style details, as shown on the website.
`description`	`string`	Marketing description text.
`category_tree`	`list[string]`	Hierarchical category path (breadcrumbs) starting with `"Products"`.
`style`	`string`	Placeholder, currently left blank.
`image_urls`	`list[string]`	List of image file paths (local filenames if downloaded). In this dataset, only the first image per product is included.
`source_url`	`string`	Canonical IKEA US product page URL.
`price`	`string`	Numeric price string (USD).
`good_to_know`	`list[string]`	Bullet-point notes from IKEA's "Good to know" section.
`materials`	`list[string]`	List of materials used, extracted from the product page.
`care_instructions`	`list[string]`	List of care instruction strings.

Example Entry

json

{
  "product_id": "00481688",
  "title": "ÅBROTTEN foam mattress, firm/white, Queen - IKEA US",
  "description": "ÅBROTTEN foam mattress, firm/white, Queen A 7⅞\" firm mattress with 3 layers of comfort and support...",
  "category_tree": ["Products", "Beds & mattresses", "Mattresses", "Foam mattresses"],
  "style": "",
  "image_urls": ["images-us/abrotten-foam-mattress-firm-white__1142853_pe881374_s5.jpg"],
  "source_url": "https://www.ikea.com/us/en/p/abrotten-foam-mattress-firm-white-00481688/",
  "price": "399",
  "good_to_know": ["To be completed with a slatted bed base...", "..."],
  "materials": ["Ticking/ Ticking, side: 100 % polyester (100% recycled)", "..."],
  "care_instructions": ["Mattress cover: Machine wash warm...", "..."]
}

Data Collection Process

The dataset was created with the following pipeline:

1.Scraper (`ikea_webscraper.py`)

—Crawls the official IKEA sitemap (sitemap.xml) filtered for US market pages.
—Extracts product metadata: titles, descriptions, categories, price, materials, care instructions, etc.
—Writes structured JSON lines to products.jsonl.

1.Image Downloader

—image_downloader_first.py: Downloads only the first "hero" image per product. → This was used to create the dataset you see here (images-us/).
—image_downloader_all.py: Alternative script to download all available product images. → Useful if you want richer visual coverage at the cost of larger storage.

1.Utilities (`image_utils.py`)

—Extracts image URLs from structured data (<script type="application/ld+json">) or OpenGraph tags.

Size

—\~30,000 products
—Metadata JSONL: \~50 MB
—Images (first only): \~5–8 GB
—Images (all): \~30–40 GB (if you run image_downloader_all.py)

Usage

You can load the dataset with Hugging Face Datasets:

python

from datasets import load_dataset

dataset = load_dataset("jeffreyszhou/ikea-us-products-2025", split="train")
print(dataset[0])

Images will be available as local files under images-us/.

Limitations & Notes

—Time-bounded: Snapshot of IKEA US products as of July 2025. Future availability or pricing may differ.
—Images: This release includes only the first product image. Run image_downloader_all.py to fetch all images if desired.
—Data quality: Parsing depends on IKEA’s frontend structure at the time of scraping. Fields may be missing or incomplete for some products.
—Style field: Currently unused.
—Errors: Failed downloads are logged in data/failed_products.txt.

License & Disclaimer

⚠️ Important: Product descriptions, images, and trademarks belong to IKEA. This dataset is provided strictly for research and non-commercial purposes (e.g. NLP, computer vision, information retrieval).

If you plan to use the dataset for anything beyond academic or hobbyist research, consult IKEA’s licensing policies and terms of service.

Citation

If you use this dataset, please cite it as:

@misc{ikea_us_products_2025,
  title = {IKEA US Product Dataset (July 2025)},
  author = {Jeffrey Zhou},
  year = {2025},
  url = {https://huggingface.co/datasets/jeffreyszhou/ikea-us-products-2025}
}