How to Download Property Data (CSV & Parquet)

You can download the entire United States in one file, or narrow to a state, county, or ZIP code. Smaller areas mean smaller files — a single ZIP is a few megabytes; the whole country is tens of gigabytes. Browse by state and drill down to the area you need.

• Parquet — the complete file, including nested price_history and tax_history. Columnar and compressed, so it's smaller and fast to query. Best for analysis in pandas, DuckDB, Polars or Spark.
• CSV — a flat table (no history columns) that opens in Excel, Google Sheets, or any tool. Best for a quick look or non-technical use.

Full field list is in the data dictionary. Everything is CC0 — free for any use.

Parquet reads in one line most places:

import pandas as pd
df = pd.read_parquet("realestatedataset-florida-2026-07-23.parquet")
print(df.shape)

Or query it directly with DuckDB — no load step, no database:

SELECT city, beds, last_sold_price
FROM 'realestatedataset-florida-2026-07-23.parquet'
WHERE beds >= 3 AND status = 'active';

CSV opens anywhere, including a spreadsheet.

A downloaded file is a snapshot — accurate the day it was compiled (the date is in the filename) and slowly aging after that. If you need current data, or you'd rather query one property than host the whole file, every row's id is a live endpoint on Straply's API. Same records, always fresh.