Malaysia Real Estate Market Data Delivered on Schedule

The client needed a continuous read on what had just come onto the Malaysian property market. The requirement was simple to state and less simple to run: every week, a dataset of the real estate listings that had appeared for sale on PropertyGuru.com.my during the preceding seven days.

Sources: propertyguru.com.my
Coverage: properties for sale in six neighbourhoods of Kuala Lumpur and Selangor, with every unique detail from the listing, including realtor and agency information
Frequency: weekly, every Monday, from January 2024 to July 2025
Volume: a full base of about 30,000 listings, then a weekly file of new listings and listings with a changed price
Delivery: CSV and XLSX files on FTP
Setup and launch: 2 working days to an approved sample

A rolling seven-day window, not a one-time dump

Scraping a site once is a bounded task. Scraping it every week against a seven-day window is a different kind of job: the target changes between runs, and the value of each delivery depends on the window being correct. Anything older than seven days was noise for this client; anything posted inside the window but missed could not be repaired afterwards, because by the next run it was no longer new. The first full delivery was the whole base. After that, each Monday file carried only the difference from the week before: listings that were new, and listings whose price had moved.

Full listing detail, realtors and agencies included

The specification asked for all unique details from each listing rather than a shortlist of fields, including the realtor and the agency behind it: the people and companies attached to a property are part of the record, not metadata to be dropped along the way.

The client's list ran to 25 fields, from floor size, tenure and Bumi lot status to the agency registration number, the photos and everything under unit features. After the first file they asked for three more - the agent's phone number, the total number of units and the developer - and these were in the data from the next full run.

Cloudflare Turnstile on the target site

PropertyGuru.com.my is protected by Cloudflare Turnstile, the kind of detail that quietly stops an in-house script. For us it is routine work, handled inside the collection layer rather than in the client's data.

It was not free either. Getting the first full collection through took about a week of tuning in January 2024; the setup that worked was a real browser session, with a 98% success rate over the first 8,900 production requests. In March 2025 the check tightened and started looping, letting through only about 100 pages a day. The session handling was rebuilt, and that week's file still went out on the Monday.

Two working days to launch a weekly Malaysia real estate feed

Payment arrived on 3 January 2024, the sample was on the FTP the next day, and the client approved it on 5 January - two working days from start to a confirmed format. The first production file, on 12 January, carried about 25,000 rows, each with the listing's own details together with its realtor and agency; the full base of 30,000 followed later that month. The result was a repeatable weekly picture of new supply, delivered without anyone on the client's side having to ask for it.

19 months of Malaysia real estate data, four files rebuilt from saved pages

Between 29 January and 5 February 2024 PropertyGuru changed the markup of the address field. The Monday files for 5, 12, 19 and 26 February went out without addresses. A fresh crawl would only have shown the listings as they stood by then. It was not needed: the scraper had kept every page it collected since 29 January. We fixed the parser, rebuilt the four February files from the saved pages, then the 29 January file as well, and all of them were back on the FTP with addresses within two days.

The feed ran for 19 months, until the end of July 2025, and the files stayed on the FTP for the client until the end of August.

See our PropertyGuru scraper and more real estate data projects; data can also come through our API or as a JSON feed. To set up a weekly feed for your own neighbourhoods, contact our team.