--- title: "UAE Data glossary" description: "Plain-English definitions of every metric and field in UAE Data: how listings are collected and de-duplicated, how prices, rents, yields, confidence, comparables, price cuts, forecasts and valuatio..." canonical: "https://uae.data.evoost.ai/glossary" last_updated: "2026-10-26" --- # UAE Data glossary Plain-English definitions of every metric and field in UAE Data: how listings are collected and de-duplicated, how prices, rents, yields, confidence, comparables, price cuts, forecasts and valuations are computed, and how plans, API keys, MCP, exports and webhooks work. ## Listings and sources ### Listing An advert for a property for sale or for rent. UAE Data merges the adverts for the same property across portals and agents into one canonical listing with its price history, so each property is counted once. (https://uae.data.evoost.ai/glossary#listing) ### Canonical ID `canonical_id / id` — The identifier of a canonical (de-duplicated) listing. Use it with GET /v1/listings/{id}, POST /v1/listings/batch and the dashboard listing page; valuation comparables return it as canonical_id. Each portal advert that maps to it is kept as a separate source record. (https://uae.data.evoost.ai/glossary#canonical-id) ### Source `source` — Where a listing was collected: major UAE property portals (such as Property Finder, Bayut and dubizzle), brokerage websites and developer websites. One canonical listing can have several sources. (https://uae.data.evoost.ai/glossary#source) ### De-duplication across portals Adverts for the same unit published on different portals or by different agents are matched by a listing signature and merged into one canonical listing, so counts and statistics don't double-count a property. (https://uae.data.evoost.ai/glossary#dedupe) ### Freshness and last seen `last_seen_at, generated_at` — Sources are re-crawled continuously, and every source record keeps when our pipeline last saw it; a listing's liveness is the most recent of those. An active listing that its still-active sources have not re-observed for 60 days is marked stale (reversed if it is seen again). Listings whose only sources are paused are not marked stale, because that gap is ours, not the market's. Analytics responses carry generated_at. (https://uae.data.evoost.ai/glossary#freshness) ### Quarantine and data quality Records that fail ingestion quality checks are held in quarantine instead of reaching the API. Accepted records are sanitised (invalid coordinates, non-positive sizes and implausible rents are blanked rather than published) and quality-scored, and analytics apply their own plausibility filters, such as hiding yields outside 1-12%. The trust center reports quarantined rows in the last 24 hours. (https://uae.data.evoost.ai/glossary#quarantine) ### Coverage Dubai and Abu Dhabi, residential and commercial listings for sale and for rent, plus off-plan projects and Dubai Land Department registries. Listing price history starts on 11 February 2026. (https://uae.data.evoost.ai/glossary#coverage) ### Emirate `emirate` — The top level of location. UAE Data covers Dubai and Abu Dhabi; filters accept exactly those two values. (https://uae.data.evoost.ai/glossary#emirate) ### Area, community and building `area, building` — Locations go from emirate to area (a community or district such as Dubai Marina or Jumeirah Village Circle) to building or tower, with a sub-community where the source gives one. Statistics are segmented by emirate, area, property type and bedrooms; area pages are addressed by a slug of area and emirate. (https://uae.data.evoost.ai/glossary#location-hierarchy) ### Property type `property_type` — Every listing is classified into one canonical type. Residential: Apartment, Studio, Penthouse, Duplex, Hotel Apartment, Townhouse, Villa, House and Whole Building. Non-residential: Office, Retail, Warehouse and Plot. (https://uae.data.evoost.ai/glossary#property-type) ### Bedrooms `bedrooms` — Bedroom count, where 0 means a studio and an empty value means unknown. Listings with an unknown bedroom count are left out of bedroom-level statistics such as benchmarks. (https://uae.data.evoost.ai/glossary#bedrooms) ### Off-plan `offplan` — Property sold before construction is complete. Off-plan projects (from developers' new-project feeds) carry the developer, completion status, handover date, payment plan, unit types, starting price and construction progress; listing search can filter off-plan units. (https://uae.data.evoost.ai/glossary#off-plan) ### Developer (DLD registry) Developers registered with the Dubai Land Department, from DLD open data, with their registration and licence details (/v1/registry/developers). Registry names are legal names and can differ from the brand names used in listings. Registry data is published as-is and not mixed into listing statistics. (https://uae.data.evoost.ai/glossary#developer) ## Prices, rents and units ### Asking price vs transaction price Prices and rents in UAE Data are advertised asking prices from live listings, not completed sale or lease transactions. Closing prices can differ from asking prices. Dubai Land Department transaction prices are not part of the dataset. (https://uae.data.evoost.ai/glossary#asking-vs-transaction) ### Annualised rent `annual rent` — Rents are compared as a yearly figure: daily × 365, weekly × 52, monthly × 12, yearly × 1. Annualised rents outside AED 10,000-10,000,000, or with an unknown or implausible frequency, are excluded from rent statistics. (https://uae.data.evoost.ai/glossary#annualized-rent) ### AED/m² and AED/ft² `price_per_sqm, ppsqm_aed` — Price divided by built-up area. The API normalises sizes to square metres (sizes given in square feet are converted at 1 ft² = 0.092903 m²) and returns AED per m². For AED per ft², divide AED/m² by 10.764. For rents, AED/m² is per year. (https://uae.data.evoost.ai/glossary#price-per-area) ## Statistics and confidence ### Median, p25 and p75 `median, p25, p75` — The median is the middle value of a sorted sample; p25 and p75 are the 25th and 75th percentiles, so half of the comparables lie between them. Medians are used instead of averages because a few extreme adverts barely move them. (https://uae.data.evoost.ai/glossary#median) ### Interquartile range (IQR) p75 minus p25: the spread of the middle half of a sample. (https://uae.data.evoost.ai/glossary#iqr) ### Outliers trimmed `outliers_removed` — Segment statistics drop values more than 3 × IQR below p25 or above p75 (a Tukey-style fence, applied when a sample has at least 8 values). Benchmarks also drop listings with an unknown bedroom count and rents mislabelled as sales. (https://uae.data.evoost.ai/glossary#outliers) ### Sample size `sample_size, n_sale, n_rent` — The number of listings behind a statistic. It is returned with every yield, benchmark, forecast and valuation so you can tell a strong signal from a thin one. (https://uae.data.evoost.ai/glossary#sample-size) ### Confidence level `confidence: high | medium | low` — A label for how much data supports a figure. Yields and area statistics: high with at least 20 sale and 20 rent listings in the segment, medium with at least 10 each, otherwise low. Benchmarks: high with at least 30 comparables, medium with at least 10. Opportunity scores: high with at least 15 sized sale and 10 rent comparables, medium with at least 8 and 5. Forecasts: high with at least 12 weeks, 100 samples and R² ≥ 0.5, medium with R² ≥ 0.25. Valuations report their own confidence together with the comparable scope. (https://uae.data.evoost.ai/glossary#confidence) ### Gross yield `gross_yield_pct` — Median annual asking rent divided by median asking sale price, × 100, for the same emirate, area, property type and bedroom count. A segment needs at least 5 sale and 5 rent listings, and yields outside 1-12% are hidden as likely data errors. Gross means before service charges, fees, vacancy and financing. (https://uae.data.evoost.ai/glossary#gross-yield) ### Net yield Not computed by the API. The free ROI calculator estimates it in your browser: annual rent after vacancy, minus service charges and maintenance, divided by the purchase price. (https://uae.data.evoost.ai/glossary#net-yield) ### Comparables and similarity `comparables, similarity` — Comparables are active listings like the property being priced. Benchmarks use the same emirate, area, property type and bedroom count. Valuations give each comparable a similarity score (higher means closer to the property being valued) and list the closest first. (https://uae.data.evoost.ai/glossary#comparables) ### Scope fallback `scope` — When there are too few close comparables, a valuation widens its search in steps and reports which one it used: building (same building), area_bedrooms (same area, type and bedrooms), area_type (same area and type) and emirate_type (same type across the emirate). Wider scopes are less precise. (https://uae.data.evoost.ai/glossary#scope-fallback) ### Benchmark verdict `position, percentile, vs_median_pct` — Where a price sits among its comparables: below_market when it is under p25, above_market when it is over p75, otherwise at_market. The percentile is the share of comparables priced at or below it, and vs_median_pct the difference from the median. (https://uae.data.evoost.ai/glossary#benchmark-verdict) ## Market signals ### Price cut and price-cut share `share_reduced_pct` — A listing counts as reduced when its current asking price is 1-30% below the first asking price our pipeline observed; moves below -30% or above +100% are excluded as likely data errors. The price-cut share is reduced listings divided by the segment's active listings minus those excluded as implausible, × 100. (https://uae.data.evoost.ai/glossary#price-cut) ### Days on market `days_on_market` — Days since our pipeline first observed the listing, not the portal's original listing date. (https://uae.data.evoost.ai/glossary#days-on-market) ### Market velocity Daily counts of new listings, price updates and listings that went inactive (/v1/analytics/market, Starter plan and up), a gauge of how fast a market is moving. (https://uae.data.evoost.ai/glossary#listing-velocity) ### Price trend (90 days) `price_trend_90d_pct` — In a valuation's market context, the percentage change in asking prices for the segment over the last 90 days. Separately, /v1/analytics/price-trends counts asking-price cuts and increases on active listings over a chosen window per emirate and property type. (https://uae.data.evoost.ai/glossary#price-trend-90d) ### Forecast `forecast, direction` — A directional outlook: a linear trend fitted to weekly median asking prices over the last 180 days, projected over the requested weeks with an approximate 95% band. It is inconclusive, with no projection, when there are fewer than 6 weeks or 30 samples, the fit is weak (R² < 0.25) or the trend is implausibly steep (more than 2% a week). Not advice. (https://uae.data.evoost.ai/glossary#forecast) ### Opportunity score `score, score_breakdown` — A 0-100 ranking of below-market listings: value (discount vs fair value, saturating at 30%, max 40) + yield (estimated gross yield from 4% to 9%, max 30) + negotiation (price cuts, max 15) + time on market (max 10), multiplied by confidence (high 1.0, medium 0.85, low 0.65). Likely data errors (discount over 40%, AED/m² under 45% of comparables, yield over 12%, rentals mislabelled as sales, unknown bedrooms) are excluded by default. (https://uae.data.evoost.ai/glossary#opportunity-score) ## Valuation ### Property valuation (AVM) `GET /v1/valuation` — An automated valuation model built on comparable asking-price listings: a sale estimate with a low-high range, an annual and monthly rent estimate, gross yield, confidence, the comparable scope used and the comparables themselves, plus area market context. It is a statistical estimate on asking prices, not a formal or RICS valuation and not based on closed DLD transactions. Available on every plan (Free shows up to 3 comparables per side). (https://uae.data.evoost.ai/glossary#valuation-avm) ### Price check `price_check` — When an asking price or rent is supplied, a valuation places it below, at or above its estimate, with the percentage difference and its percentile among the comparables. (https://uae.data.evoost.ai/glossary#price-check) ### Dispersion `dispersion_pct` — How spread out the comparables' prices per m² are: their interquartile range divided by their median, as a percentage. Low dispersion means the comparables agree; high dispersion means a mixed set and a less certain estimate. It feeds the valuation's confidence level. (https://uae.data.evoost.ai/glossary#dispersion) ### FSD (forecast standard deviation) `fsd_pct` — The expected statistical error of a valuation estimate, as a percentage of the estimate. The lower the FSD, the tighter the estimate: in AVM practice, roughly two in three outcomes are expected within ± one FSD of the estimate. It describes the asking-price model, not the price a property will sell for. (https://uae.data.evoost.ai/glossary#fsd) ### Size tolerance `size_tolerance_pct` — How far a comparable's size may differ from the property being valued, as ± a percentage of its size (from 5% to 50%, default 30%). A narrower tolerance gives closer comparables but may find fewer of them. (https://uae.data.evoost.ai/glossary#size-tolerance) ### Asset lookup `GET /v1/properties/lookup` — Search by name, listing URL or listing id for a building, project, area or listing, with its area, emirate, typical sizes (p25, median, p75), bedroom mix and listing count, to prefill a valuation. A listing found this way can be valued directly by its canonical_id. (https://uae.data.evoost.ai/glossary#asset-lookup) ### Market rating `market.rating` — A 1-5 score for the area and segment of a valuation, combining the asking-price trend, liquidity (how quickly listings turn over), price dispersion and price cuts; its components are returned as rating_components. It summarises listing activity, not transactions, and is not investment advice. (https://uae.data.evoost.ai/glossary#market-rating) ### Market cycle `market.cycle` — Whether asking prices per m² in the area are rising, stable or declining over recent months, or unknown when there is not enough history. (https://uae.data.evoost.ai/glossary#market-cycle) ### Market risk `market.risk` — A low, medium or high label for how uncertain pricing is in the area, from the same signals as the market rating (trend, liquidity, dispersion and price cuts). (https://uae.data.evoost.ai/glossary#market-risk) ### Withdrawn listings `market.withdrawn_90d` — Listings that disappeared from the portals in the last 90 days, with their median size and price per m². Some were sold or let, others were withdrawn or relisted elsewhere, so they are never reported as sales or DLD transactions. (https://uae.data.evoost.ai/glossary#withdrawn-listings) ### Portfolio (batch) valuation `POST /v1/valuation/batch` — Values many properties in one call, each with the same parameters and result as a single valuation and its own ok/error status: up to 25 items on Pro and 250 on Enterprise. Each item counts as one request. (https://uae.data.evoost.ai/glossary#portfolio-valuation) ## Plans and access ### Plans and quotas Each account has a plan with a monthly request quota, a per-minute rate limit and a set of features: Free 1,000 requests a month (30 a minute), Starter 20,000 (60), Pro 150,000 (120) and Enterprise custom. Higher plans unlock more endpoints; a gated endpoint answers 403 plan_upgrade_required with the required_plan. The live matrix is on the pricing page. (https://uae.data.evoost.ai/glossary#plans) ### API key `x-api-key` — A secret starting with uae_live_, sent in the x-api-key header. It is shown once at creation and stored only as a SHA-256 hash; rotate or revoke it from the dashboard. Dashboard keys carry the read scope. (https://uae.data.evoost.ai/glossary#api-key) ### Rate limits `RateLimit, retry-after` — A per-minute request limit per account set by the plan. Responses carry RateLimit, RateLimit-Policy, x-ratelimit-limit and x-ratelimit-remaining headers; going over answers HTTP 429 rate_limited with retry-after, and an exhausted monthly quota answers 429 quota_exceeded. (https://uae.data.evoost.ai/glossary#rate-limits) ### MCP server UAE Data's Model Context Protocol server at /mcp lets AI agents search listings, read yields, benchmarks, forecasts, price cuts and off-plan data, value a property and start exports, as the signed-in user. A separate public documentation MCP server at /docs/mcp serves the docs only. (https://uae.data.evoost.ai/glossary#mcp) ### OAuth connector for personal agents Claude, ChatGPT, Cursor, VS Code and other MCP clients connect to /mcp and ask you to sign in to UAE Data (OAuth authorization code with PKCE). The agent then acts as you, with your plan and quota; identity is never taken from tool input. Disconnect the app in your client to stop it. (https://uae.data.evoost.ai/glossary#oauth-connector) ### Exports `format=csv, POST /v1/exports` — CSV export (Pro and up): add format=csv to listings search, yields, areas, price cuts and off-plan. Async bulk exports (Pro and up) queue a CSV or JSONL file of up to 25,000 rows on Pro or 250,000 on Enterprise, kept for 7 days, with at most 2 jobs running at once. (https://uae.data.evoost.ai/glossary#exports) ### Webhooks (market alerts) `/v1/webhooks` — Real-time market alerts through the API (Pro and up: 5 subscriptions on Pro, 50 on Enterprise): an https endpoint subscribes to listing.created and listing.price_changed events that match listing-search filters and receives batched deliveries signed per the Standard Webhooks spec, in JSON or Slack format. There is no dashboard screen for webhooks yet. (https://uae.data.evoost.ai/glossary#webhooks) Methodology and data freshness: https://uae.data.evoost.ai/trust.md. API reference: https://uae.data.evoost.ai/docs.md.