Measured 2026-09-15T06:44:21.205199-04:00. The homepage refreshes after a successful nightly measurement. If a source or measurement fails, the last successful figures and date stay visible.
2,738,421 Florida parcel records and 17,581,928 listing image records. Parcels are not a count of individual dwelling units. Image records are not deduplicated photographs.
| County | Parcel records |
|---|---|
| 11 | 1 |
| 15 | 1 |
| Broward | 754,371 |
| Miami-Dade | 939,461 |
| 28 | 2 |
| 38 | 2 |
| Indian River | 95,446 |
| 46 | 3 |
| Martin | 98,197 |
| 57 | 3 |
| 58 | 2 |
| Palm Beach | 658,342 |
| St. Lucie | 192,590 |
762,044 sale records; 564,757 rental records; 1,891 other closed records. Each listing key is counted once. A property can have multiple listings over time. Business opportunities and unrecognized categories are kept separate.
| Listing category | Counted as | Records |
|---|---|---|
| Business Opportunity | other | 1,891 |
| Commercial Land | sales | 4,000 |
| Commercial Sale | sales | 18,024 |
| Land/Boat Docks | sales | 17,783 |
| Residential | sales | 710,189 |
| Residential Income | sales | 12,048 |
| Residential Lease | rentals | 564,757 |
Approx. 639 million. This is an estimate of populated source fields, not a count of unique real-world facts. It includes Florida company data as well as property datasets; it is not limited to the three counties our brokerage serves.
We count rows exactly in six reviewed primary datasets. For each, we estimate populated fields per row from three repeatable samples spread across database blocks, targeting 3,000 rows per sample. We multiply each exact row count by its pooled sample average. The measurement uses one consistent database snapshot.
Nulls, whitespace-only text and empty containers do not count. Zero and false values do count. A populated JSON field counts once, including arrays and nested objects. We exclude working copies, derived tables, photo metadata, source labels, ingestion timestamps and specified technical identifiers. Normalized listing columns are not added again to the original listing payload.
Independent records can repeat a fact, and source fields may overlap in meaning. This method removes duplicate database copies; it does not claim semantic deduplication across every source. Unreviewed datasets are omitted. Changes in the estimate can reflect ingestion, sampling or the method, not just growth.
| Primary dataset | Exact rows | Estimated populated fields |
|---|---|---|
| Florida parcel tax rolls | 2,738,421 | 127,542,708 |
| Broward recorded documents | 2,086,704 | 9,737,175 |
| Closed MLS listing records | 1,328,692 | 163,669,576 |
| Property permits | 4,339,191 | 37,472,831 |
| Florida company records | 12,818,747 | 290,917,007 |
| Municipal code cases | 1,755,421 | 9,183,339 |
The three sample-based total estimates ranged from 637,382,102 to 639,621,960. This is a sampling diagnostic, not a statistical confidence interval.
Method version 2 replaces the former approximately 2.5-billion estimate, which included working copies, tracking fields and empty JSON keys. The lower total reflects stricter counting, not lost data.
Download the aggregate measurement and field-selection evidence (JSON)