How our market figures are calculated

Where the figures come from

Sale prices come from the Dubai Land Department's register of sale transactions. Rents come from Ejari, the register of tenancy contracts. Net yields also use the service-charge budgets owners' associations file with the Land Department. These are the sources today. We plan to add more, and each figure will name the sources it used.

We rebuild the figures every night after the registers update. A figure only changes when a new build is published.

Price comparison

The percentage beside a sale compares its price per m² with the median of similar sales: the same completion status (off-plan or ready), property type, bedroom count and size band, registered in the same project over the last 12 months.

If the project registered fewer than 30 such sales in 12 months, we look at 24 months. If that is still too few, or the sale names no project, we compare with the whole area over 12 months. We never mix these levels, and each comparison says which one it used.

Size bands are fixed ranges, each 1.5 times as wide as the one below, such as 58–86 m² and 86–130 m². A sale is compared with the other sales in its band, so one near a band's edge can be compared with properties up to about 50% larger or smaller.

A sale gets no comparison when its size or price fails our data checks, when it is outside the 12 or 24 month window, when fewer than 30 similar sales remain, or when its own price per m² is far outside the similar sales. We do not then try a wider set, because that would let the answer choose the comparison. Rent contracts get no comparison.

Rental yield

Gross rental yield is the median yearly rent of new tenancies divided by the median price of ready sales, for the same area or project, property type and bedroom count, over the same four quarters. Renewals are left out, because a renewal restates a sitting tenant's rent. Off-plan sales are left out, because a home that has not been handed over earns no rent.

Net yield takes off one year of the project's residential service charge, using the latest budget year up to the end of the window, and the figure names the budget year it used. It covers service charges only. Vacancy, maintenance, letting fees and anything else are not taken off. Areas have no net yield, because service charges are set per project.

We publish a yield only when both sides kept at least 30 records, the Ejari register is complete for every quarter of the window, and the yield falls between 1% and 15%. Otherwise the page says which condition failed.

Across the yields we published up to October 2026, gross yields ran from about 2% for the most expensive homes to about 14% for the cheapest flats. A yield outside 1% to 15% almost always means the two medians describe different homes. The usual cause is a window where most of the sales are shares of a flat, sold at a fraction of its price. We do not cut such a figure down to the limit. We leave it out and show the median rent and price it came from.

The yield appears in the market summary on the search page when your filters pick one area or project, one bedroom count, and units or villas. With no property type chosen, it is for units.

Price change

Price change compares the median price per m² of sales in the last four quarters with the same four quarters 1, 3 and 5 years earlier. Off-plan and ready sales are measured apart. For 3 and 5 years we also give the yearly rate that compounds to the total.

Area pages can also show the last quarter on its own, against the same quarter 1, 3 and 5 years earlier. A quarter holds fewer sales than a year, so its figures move more. Projects are measured over a full year only, because one project rarely registers 30 sales of a kind in a quarter.

These are medians of what sold, not the same homes sold twice. The registers do not link a sale to the next sale of the same home. A change can come from a different mix of homes selling as much as from prices moving. Using price per m² keeps a shift towards larger homes from showing as growth.

A change is published only when both periods kept at least 30 sales. Otherwise the cell says whether the period now or the period then held too few sales, or that there were no sales then to compare with.

Up-and-coming areas

An up-and-coming area passes three tests on its ready sales, for units or for villas, every size together. We start from the areas with a published price change over 1 year, the measured areas.

First, its median price per sqft over the last four quarters is at or below the middle measured area's. With an even number of areas the middle is the lower of the two middle values. Second, its price change over 3 years is published and above zero. Third, it kept at least as many sales in the last four quarters as in the same four quarters 1 year earlier. Equal counts pass.

Areas that pass are ranked by the yearly rate of their change over 3 years, which moves less with the mix of properties sold than a shorter change does. We make a list only when at least 10 areas were measured. With fewer, the cheaper half is a handful of areas, so we give the count and no list.

The rule picks areas on what has already happened. It is not a forecast or a score. Off-plan sales and projects are left out.

Data checks

Before taking a median we leave out records whose floor area is under 150 sq ft, records that cover several properties at one price, and records whose price per m² is far outside the rest of their group (3 times the spread of the middle half, measured on a log scale). Every figure counts what it kept and what it left out. We never fill gaps or smooth figures.

We also leave out a sale whose floor area is under 269 sqft for each bedroom it was registered with. DLD records the sale of a share in a property as a sale of the whole property’s bedrooms with the share’s part of the floor area, at a fraction of the price. The register does not say which sales are shares, so size is the only sign. These sales stay listed, marked “Possible share sale”. Rents are not held to this check.

Rent coverage

Ejari contracts are often registered weeks after they start, so a quarter keeps filling after it ends. We use a quarter for rents only once it has had time to settle and holds a normal number of contracts. Rent figures stop at the last quarter in an unbroken run of complete quarters. Today that is 31 Mar 2026.