← Catalog

Airbnb Dataset Sample: Free CSV Data Download

You want to see the Airbnb data before you commit to it. Fair. This page hands you the sample first: 20 real listings as CSV and JSON, free, no signup, with all 12 columns of the listings-summary schema intact. Test your import script on it, check the column names against your model, plot the coordinates on a map. When the sample proves the schema fits, the full 90,169-row pack across 6 US cities is one click away, and the MCP server lets you query it without downloading anything.

Download the Airbnb CSV sample

Two files, same 20 rows from Austin:

sample.csv: the rows as plain CSV, opens in Excel, Sheets, pandas, or a text editor
sample.json: the same rows as JSON, with the column list on top, handy for API mocks and test fixtures

Or pull it straight into a script:

curl -s -o airbnb-sample.csv https://jayjex.github.io/data-vault/data/airbnb-six-cities/sample.csv

The first rows look like this:

id,name,neighbourhood,latitude,longitude,room_type,price,...
934859141300145599,An Exquisite Rooftop Studio,78757,30.3636,-97.7262,Entire home/apt,61.00,...
944156894640610027,Tarrytown bungalow,78703,30.2988,-97.7803,Private room,149.00,...

More free samples like this one, across five datasets: sample data downloads.

The 12 columns in the sample

Every row carries the full listings-summary schema, 12 columns, no host identity fields:

id, name, neighbourhood, latitude, longitude, room_type, price, minimum_nights, availability_365, number_of_reviews, reviews_per_month, last_review

A few worth knowing before you build on them:

neighbourhood holds a ZIP code area, so 78701 filters downtown Austin directly.
latitude and longitude are blurred by about 11 meters upstream at Inside Airbnb, enough to map neighbourhoods without exposing exact addresses.
price is the nightly rate in USD, cleaned from the source text.
availability_365 counts bookable days over the next 365 days from the snapshot date, a quick proxy for how active a listing is.
reviews_per_month and last_review can be empty for listings with no reviews yet, so handle nulls in your loader.

The Airbnb data guide explains every column in detail, including what each room type covers.

The dataset behind the sample

The 20 sample rows are the head of a 90,169-row pack: one listings file per city, same 12 columns throughout, captured from Inside Airbnb on the dates below and written here as-is:

City	Snapshot date	Listings
Austin, TX	2026-06-22	11,295
Nashville, TN	2026-06-26	10,242
Denver, CO	2026-06-30	4,939
Las Vegas (Clark County), NV	2026-06-27	20,234
New York City, NY	2026-08-10	30,234
San Diego, CA	2026-06-27	13,225

A per-city index with room-type mix ships alongside as cities-index.csv. Dataset-level stats and the preview table live on the Airbnb Listings, 6 US Cities dataset page.

Query the full data before you download it

The sample proves the schema, but 20 rows answer no real question. The dataset-mcp server covers that: it serves read-only queries against the complete 90,169-row files over the MCP protocol, so an agent or script can filter and paginate without pulling the whole pack.

npx --allow-git=all -y github:jayjex/dataset-mcp

Five listings in downtown Austin, by ZIP:

query_dataset("airbnb-six-cities", {where: [{column: "neighbourhood", op: "=", value: "78701"}], limit: 5})

Column stats for a quick sanity check:

get_stats("airbnb-six-cities", {column: "price"})

Setup steps and more worked examples: For agents. If you are weighing file downloads against live calls in general, CSV vs API breaks down when each one fits.

FAQ

Is the Airbnb dataset sample free to use?

Yes. The sample.csv and sample.json files download with no signup, and they are licensed CC BY 4.0. If you republish them, credit Inside Airbnb as the source plus Data Vault, and the same terms apply to the full 90,169-row pack.

What is inside the Airbnb CSV sample?

Twenty real listings from Austin with all 12 columns of the listings-summary schema. The full pack uses the same 12 columns for every city, so code written against the sample runs unchanged on all six markets in the table above.

Can I check the full 90,169-row dataset before downloading it?

Yes. The dataset-mcp server answers read-only queries against the complete files, with =, contains, gt, and lt operators plus column stats. The query examples above show the shape.

Can I compare these nightly prices to HUD fair market rents?

Careful with that comparison. A nightly short-term rate and a monthly FMR measure different markets; FMR vs actual rent shows why the two numbers sit apart and what to match up before benchmarking one against the other.

Get the full pack

Download free sample (CSV) Get the full pack

Full pack: 90,169 rows across 6 cities, per-city CSVs, data dictionary, methodology, source checksums. Checkout and download run through Getly.