Taiwan Real Estate Actual-Price Transactions - Per Record

Taiwan's actual-price registration transactions - the official record of real-estate deals - as one clean, flat record per transaction, for the quarter you choose. Sale, presale and lease transactions, each with district, address, area, price, date and building attributes. Chinese text delivered verbatim; no buyer or seller name.

What one record represents

The source is the Taiwan Ministry of the Interior Real Estate Actual Price Registration (plvr.land.moi.gov.tw), downloaded as the ministry's quarterly ZIP for the season you select. Each record is one transaction from the main tables: its district, transaction target, land+building address, land area, transaction date (ROC calendar), building area, room/hall/bath counts, building attributes, total price and unit price, parking details and a serial number - plus a transaction_type (sale, presale or lease) and the county code.

Coverage and volume

For season 115S2 the release holds 131,580 transactions - 75,399 sales, 8,571 presales and 47,610 leases, measured at build time from the downloaded ZIP.

Sol's Wave-4 index put this door at 131,580 EXACT for 115S2; measured live at build time the season's three main tables sum to exactly 131,580 (75,399 + 8,571 + 47,610) - matching the index.

Each quarter is one partition: set the Season input (ROC year + quarter, e.g. 115S2) to pull that release. The Actor downloads the season ZIP, iterates the per-county sale/presale/lease main tables (skipping their two header rows), and delivers each transaction, stopping as soon as your Maximum records cap is met.

Licence and attribution

The data is released under the Open Government Data License, Taiwan, version 1.0 (OGDL 1.0). The full notice travels on every record:

Open Government Data License, Taiwan, version 1.0 (OGDL 1.0): the data may be used for any purpose, including commercial products and services, provided the source (Ministry of the Interior) is noted.

Person-data policy

Taiwan actual-price registration carries no buyer or seller name - the record is address x area x price x date. A per-record assertion rejects any buyer/seller/owner/person field (verified with a planted-field test). No natural-person data is processed.

Interpretation caveat

One record per Taiwan actual-price registration transaction for the selected season: sale, presale and lease transactions, each carrying district, address, land/building area, transaction date, price and building attributes, plus a transaction_type and county code. Chinese text is delivered verbatim. The registration carries no buyer or seller name.

Values are reproduced verbatim from the source; the Actor never rewrites a field, and Chinese text is delivered as-is with no transliteration - relabelling would be a modification. Dates are in the ROC calendar (e.g. 1150601 = 2026-06-01). Prices are in New Taiwan Dollars; compartmented and has_management_org are delivered as real booleans.

Provenance and compliance

Every run reads the door host's robots.txt at runtime; the gate result (URL, status, byte length, SHA-256 of the policy) is written to the run's RUN_RECEIPT, and the data path is confirmed permitted before any fetch. The endpoint is keyless. The Actor never bypasses a block or fetches through a mirror.

Data quality and freshness

Boolean columns are delivered as real booleans and numeric columns as real numbers. Delivery is keyed on a stable composite id, so the dataset is safe to diff, deduplicate or upsert. Every run re-reads the live door, so the data is as fresh as the source publishes, and each record's observed_at stamp dates the snapshot. The run's RUN_RECEIPT records the source URL and how many records were delivered and charged, and confirms charge_equals_delivered.

Billing and delivery

Pricing is per record: you are billed only for records the Actor actually delivers, at a flat rate, with the charge raised after each record is pushed (push-then-charge), so a failed or empty run costs nothing. There is no subscription and no minimum. The Maximum records cap bounds every run, so you control the spend precisely - set it low to sample the shape of the data cheaply, then raise it once you have confirmed the cell fits your use. Because delivery is incremental and keyed on a stable id, re-running against the same door and upserting on that id keeps a downstream table current without re-paying for rows you already hold; only the new and changed records add cost. Every run's receipt reconciles delivered against charged, so the billing is auditable against the data you received.

Inputs

• Season (season, required) - the quarterly release as ROC year + S + quarter (1-4), e.g. 115S2.
• Maximum records (maxRecords) - hard cap on transaction records delivered and billed within the season.

Output

Records land in the Actor's default dataset and export as JSON, CSV, Excel or via the Apify API. A tabular overview surfaces district, transaction type, address, total and unit price, building area, transaction date and season.

Fields in detail

The record carries the full transaction row - district, transaction target, address, land and building area, use zoning, transaction date, floor and building attributes, room/hall/bath counts, the booleans compartmented and has_management_org, total and unit price, parking category/area/price, a note, the serial number and the main building area - plus transaction_type, county_code and season. The provenance block closes every record.

Typical uses

Real-estate analysts and valuation teams use this cell as the authoritative Taiwan actual-price series - one queryable table per quarter, filterable by district, transaction type or building attributes, with total and unit price as real numbers ready to chart or model. A group-by on district and transaction type gives median price per area; the season key lets you assemble a multi-quarter panel by running each release and concatenating. Because the registration is anonymised, the series is safe to redistribute under OGDL 1.0.

Scaling and limits

Set Maximum records low to sample or high to pull the whole season. The Actor streams the per-county CSVs inside the ZIP and delivers incrementally, so memory stays flat and you are billed only for what is delivered. To build a longer history, run the cell once per season and concatenate on the season key; each record's observed_at stamp dates the snapshot.

How the season release is delivered

The ministry publishes actual-price data quarterly, one ZIP per season, with per-county CSVs inside for each transaction type. This cell downloads the season ZIP and iterates the main tables - the sale, presale and lease files for every county - so a single run for a season captures all three transaction types across the whole country. Each source file carries two header rows (Chinese labels and their English glosses); the Actor skips both and reads by column position, which is stable across the release, so the 29 transaction columns line up correctly for every county and type. The transaction_type and county_code on each record record which file it came from.

Reading the transaction

The three transaction types answer different questions: sale is completed real-estate sales, presale is pre-completion new-build sales, and lease is rental registrations - keeping them in one dataset with a type field lets you analyse them together or filter to one. total_price_ntd and unit_price_ntd_m2 are real numbers in New Taiwan Dollars, ready to chart or model; building_area_m2 and land_area_m2 give the areas; and the room/hall/bath counts plus compartmented and has_management_org (real booleans) describe the unit. The ROC-calendar dates (e.g. 1150601) are kept verbatim rather than converted, faithful to the source.

Building a multi-quarter panel

Each run is scoped to one season, which keeps the download bounded and the output on-grain. To build a longer history, run the cell once per season and concatenate on the season key - the schema is identical across quarters, so the panel assembles without reconciliation. Because the registration is anonymised and released under OGDL 1.0, the resulting series is safe to redistribute, and each record's observed_at stamp dates the snapshot.

Sibling Actors

It sits beside the fleet's Singapore URA property-transactions cell (0% overlap - different jurisdiction). It shares its season-zip engineering with the fleet's other records Actors.