Method and sources

Everything on this site is computed from one file: the Dubai Land Department transaction register, published through Dubai Pulse open data. Nothing is an asking price, an estimate or a projection.

What is counted

1,346,049 residential sales, registered up to3 September 2026. A sale counts as residential when the register describes it as a studio, a numbered bedroom count or a penthouse. Shops, offices and land plots are excluded, because a price per square foot that mixes a shop with an apartment describes neither.

Mortgage registrations are counted separately from sales. They are not a subset of them, which is why the financed share is expressed as registrations per hundred sales rather than as a percentage of buyers.

How a sale is matched to a community

This is the step that most often goes wrong. The transaction file carries an area_id, and the community boundary file carries aCOMM_NUM. They look like the same number and they are not. Joining them directly produces roughly ninety coincidental matches whose names disagree with each other, which is how a map ends up confidently showing the wrong prices in the wrong places.

The register's own area lookup is the bridge: area_id to the lookup's municipality number, and only then to the boundary file. Where a community is better known by another name, the familiar name is shown first and the official one beneath it, because nobody searches for Al Thanyah Fifth.

Medians, not averages

Every headline figure is a median. A handful of very large sales pulls an average upward, and in a market with as much variation as Dubai's the average of a community is frequently a number that no property there actually sold for.

A community is only reported once it records at least 20 priced sales within the period being shown. Below that a single tower or a single land plot decides the answer. That threshold is the reason parts of the map are grey, and the reason longer periods fill it in:

• Over twelve months, 95 of 223 communities are reportable, from 189,931 sales.
• Over three years, 131 of 223 communities are reportable, from 555,792 sales.
• Over five years, 138 of 223 communities are reportable, from 756,137 sales.
• Over ten years, 145 of 223 communities are reportable, from 949,448 sales.

How growth is measured

Each period is compared against the equally long period immediately before it. The twelve month figure is measured against the twelve months before that, and the ten year figure against the decade before that. A short window set against a fixed distant baseline would flatter every area the further back the baseline reached, and would hide a market that has turned recently.

Large swings in recently launched communities usually reflect a change in what is being sold there rather than a repricing of the same stock. An area whose first towers complete will shift from land and off-plan studios to finished apartments, and the median moves with the mix.

What the numbers are not

A community median is not a valuation. It covers every building, age, floor and finish in that community at once, and the spread inside a single community is often wider than the gap between two communities. Nothing here tells you what a particular property is worth.

The register records what was registered. It does not capture what was paid outside the recorded price, what incentives a developer attached, or what a payment plan was worth in present value. Off-plan sales in particular are registered at the contract price, which is not the same as a cash price today.

Boundaries are simplified for drawing. They are accurate enough to identify a community and are not a cadastral record.

Corrections

Every figure carries the date of the most recent sale behind it, currently3 September 2026. A market figure with no date is a market figure that is quietly wrong, and the date above is how you check whether this has been rebuilt since you last looked.