Loading
Research · Thread 08
Listing market intelligence
Telling a brokerage where its listings stand against the market, and refusing to send the answer when it can't be trusted.
A property portal is a ranked shelf, and most brokerages read their position on it the way everyone does: opening the site and scrolling. That answers the question for one listing, on one day, for whoever happened to look. This pipeline answers it for the whole shelf, every week, without anyone opening a browser, collecting from several portals at once and scoring every listing against the market. The research question was never the collecting, it was whether an unattended weekly report could stay trustworthy enough to believe on a week nobody was watching it run.
Paper to shipped
Implementation
Five gated stages, each allowed to stop the nextCollect from several portals at once, validate against schema and duplicates, score every listing on a nine-component visibility formula, re-verify the biggest movers with a second independent fetch, then report. Validation can refuse to hand data to scoring, and scoring can refuse to hand a week to the report, on the premise that a pipeline that always produces output eventually produces confident nonsense.
Shipped
It sends the alert, not the reportWhen the quality gates fail, the job sends an alert saying the week is not trustworthy and why, rather than a degraded report with a caveat at the top. The same guard covers an empty run, so a collection that silently returns nothing raises an alarm instead of an empty PDF.
In progress
Ran weekly from June through mid-August 2026, then stoppedRan weekly to a single recipient at a Dubai brokerage. The last delivered edition covered 1,752 listings across 30 competing agencies; the last verified end-to-end run, 2026-08-14, collected 4,394 listings across eleven communities and passed every quality gate. It stopped running after that, for reasons this thread does not claim to know.