Insights

30 Free Web Datasets We Published This Month — And What They Reveal

A roundup of free web datasets from DataHarbor: one finding from each recent sample, from IKEA sofas to Madrid rents, with links to download every CSV.

DataHarbor Team
November 18, 2026
9 min read
#free datasets#sample data#web scraping#ecommerce data#csv#data roundup
30 Free Web Datasets We Published This Month — And What They Reveal

30 Free Web Datasets We Published This Month — And What They Reveal

We publish a free sample dataset almost every day. Each one is a real CSV pulled from a public web page, with a short write-up of what is inside. This post is the hub for the recent batch: free web datasets you can download without a form, plus one concrete finding from each so you can judge the data before you open the file.

A note on the count: the headline says 30, but we only state numbers we can prove. Counting the dataset files in our library, 24 samples carry a publication date between 17 October and 18 November 2026. The findings below come from a selection of those files. Every figure is calculated from the CSV you can download yourself. The browsable list lives on our sample datasets page.

Today's addition is the IKEA sofas product catalog sample, so we start there.

What our IKEA sample shows

The IKEA file holds 100 sofa listings from the public US category, collected on 16 November 2026. Each row has the product ID, name, type and cover, current price, previous price, rating, review count, colors and an online-availability flag.

Here are the figures we computed from it:

  • Median price is $1,099, the average is $1,293. The range runs from $169 (a GLOSTAD loveseat) to $5,345 (a JÄTTEBO set). The gap between median and mean tells you the catalog has a long premium tail.
  • 43 of 100 listings show a previous price. Where a regular price is shown, the average markdown is 19.5%.
  • Average rating is 3.94 out of 5 across the 99 listings that have one. One listing has no rating yet.
  • Price bands: 15 listings are under $500, 32 sit between $500 and $999, 20 between $1,000 and $1,499, 17 between $1,500 and $1,999, and 16 are $2,000 or more.
  • Product families repeat. KIVIK appears 11 times, VIMLE 10 times and JÄTTEBO 10 times, because each cover option is its own variant row.
  • Most-reviewed listing: FRIHETEN, with 1,352 reviews and a 4.1 rating.
  • Colors: beige is tagged on 29 listings, gray on 21 and black on 14.

A five-row excerpt:

name type_description price_usd previous_price_usd rating review_count
KIVIK Sofa, Grann/Bomstad black 1299 4.2 70
UPPLAND Sofa, Hakebo gray/green 679 849 3.6 182
HYLTARP Sofa, Gransel natural 1099 1249 4.1 142
KIVIK Sectional, 4-seat with chaise, Kelinge gray-turquoise 1299 3.8 157
KIVIK Sofa, Tibbleby beige/gray 799 4.1 568

The same KIVIK name sells at $799 and $1,299 depending on size and cover. That is exactly the kind of detail a flat product count hides, and why variant-level rows matter for a free ecommerce dataset.

Findings from other recent samples

Each line below links to a dataset page where you can download the CSV. The numbers are medians or counts calculated from that file.

Retail and marketplaces

Travel, property and local

Jobs, cars and content

What the numbers say across datasets

A few patterns repeat across very different sites.

Medians beat averages. In the IKEA file the mean sits about $194 above the median. In almost every price dataset a handful of premium items drag the average up. If you build a pricing model, start from medians and bands.

Discount data is patchy. Only 43 of 100 IKEA sofas show a previous price, and only 21 of 60 Blinkit items sit below MRP. A missing previous price usually means "no offer", not "missing data", so your schema should tell those two cases apart.

Salary and rating fields are often empty. Only 23 of 99 remote job postings list a minimum salary. Plan for nulls instead of dropping rows.

Brand concentration is strong. Milwaukee is 49 of 95 Home Depot power tool listings. HP is 31 of 100 Walmart laptops. For a competitor scan, that tells you whose price moves matter most.

Who uses free sample datasets, and for what

  • Pricing and category teams compare price bands and markdown depth, as in the IKEA and Home Depot files.
  • Analysts and data scientists test a model or dashboard on real rows before buying a feed.
  • Proptech and travel teams benchmark rents, nightly rates and fares against their own numbers.
  • Recruiters and HR analysts look at which postings disclose pay.
  • Founders check whether a source holds the fields they need before building around it.

For more on why this step matters, read why you should ask for a free sample dataset before you buy.

Which fields matter in a web scraping dataset

Not all columns are equal. The ones we check first:

  1. A stable ID. Product ID, listing ID or ASIN lets you track one item over time and remove duplicates.
  2. Current and previous price. Together they give you markdown depth. In the IKEA file the average markdown is 19.5% where a previous price exists.
  3. Rating and review count together. A 4.8 from three reviews is not a 4.1 from 1,352.
  4. A timestamp. Prices change, so every row carries scraped_at_utc.
  5. An absolute URL. You can verify any row by opening the page.

Common challenges with web scraping datasets

  • Anti-bot measures. Public pages often block simple requests, so collection needs rotating access and sometimes full browser rendering.
  • Pagination that does not paginate. The IKEA category page shows 24 products and loads more through a background request. Changing a page parameter in the address returned the same 24 products. We found the site's own list request and read it in larger batches instead.
  • Price tracking. If you want to watch prices over time instead of taking one snapshot, see our ecommerce price monitoring playbook.
  • Layout changes. A redesign can break a parser overnight. Checks on row counts and empty columns catch this quickly.
  • Variants and duplicates. One product family can appear many times. Decide early whether you want one row per product or per variant.
  • Data quality. Even in a clean file, 16 of 100 IKEA rows have no color tag and 57 have no previous price. Document what null means.

How often should you refresh the data?

It depends on how fast the source changes. Promotional prices on retail sites can change within days, so weekly or daily refreshes suit price monitoring. Rents, hotel rates and flights move with demand and the calendar, so daily snapshots are common. Catalog structure and brand mix change slowly, and monthly refreshes are usually enough. The IKEA offers we collected were valid for a limited window, which is a good reminder that a single snapshot only describes one day.

FAQ

Are these free web datasets really free? Yes. Each sample is a CSV you can download from its dataset page. They are small, public-data samples meant to show structure and quality.

Can I use a free sample dataset in a commercial project? The samples are for evaluation. If you need rights, volume or refreshes for production, contact us and we will scope it.

How were the datasets collected? From public web pages only, with deduplication, normalised values and a check for empty columns. No pages behind a login and no personal data.

What if my target site is not in the list? Request a free sample of your own target and we will show you what the data looks like before you commit.

Get a free sample of your own target

If you saw a column here that you would like for a different site, category or country, request a free sample of your own target. Our datasets start from $49, see pricing for details. Meanwhile, browse all the free sample CSV files and start with the IKEA sofas dataset.

Author: DataHarbor Team

Related Posts

Check out other blog posts on this topic

Is Web Scraping Legal in 2026? A Practical Guide to GDPR, CFAA and Terms of Service
Insights

Is Web Scraping Legal in 2026? A Practical Guide to GDPR, CFAA and Terms of Service

Is web scraping legal in 2026? A plain guide to public vs personal data, GDPR, the CFAA and terms of service, with a safe example dataset.

#web scraping#legality#gdpr
Oct 26, 2026
10 min read
Read More
Tired of Getting Blocked by Cloudflare, DataDome and Akamai? Here's How to Get the Data You Need Without the Headache
Insights

Tired of Getting Blocked by Cloudflare, DataDome and Akamai? Here's How to Get the Data You Need Without the Headache

Blocked by Cloudflare, DataDome or Akamai? Learn how enterprise web scraping handles WAFs, anti-bot systems and large-scale public data extraction.

#web scraping#cloudflare#datadome
Aug 8, 2026
8 min read
Read More
No-Code vs Custom Web Scraping: Which Is Right for Your Business in 2026?
Insights

No-Code vs Custom Web Scraping: Which Is Right for Your Business in 2026?

Compare no-code vs custom web scraping in 2026. Learn which data extraction approach fits your business needs, budget, scale, and technical requirements.

#no-code scraping#custom scraping#web scraping
Aug 2, 2026
8 min read
Read More

Need this data for your own targets?

Tell us the website and the fields you need — we'll send a free sample first.