Where are people actually looking?

1 of 120 places published · 181 below the 5-person floor (busiest shown)

Where are other buyers actually looking?

3 more readings, and 1 mode this map does not have — and why

Supply — see Explore: This is the same view as Explore, not a missing one. Supply is the listing count, Explore paints the listing count, and there is no second measure that would make them different pictures. Two chips producing a byte-identical page is a control that appears to do nothing, which reads as the page being broken rather than as the two being the same question.

Demand is 5 behavioural signals, weighted by how much intent each one shows, counted in distinct people and pooled so no individual can be read out of it.

Demandpeople, not events

measured, not ranked — too few published places here to sort into bands

not published — under 5 people here

circle size = demand · number inside = listings

117 places in view are drawn without a name because the names would have overlapped at this zoom — each one is still in its mark’s tooltip, and zooming in gives them room.

One place in this view clears the floor. Demand bands need at least 4 published places to sort between, so these read Measured, not ranked — the signals behind them are real and shown in full; the ranking would not be. 181 places sit below the 5-person floor and are shown grey.

1 signal not counted — and why

Video engagement: This platform owns none of its video. All 13 clips play from the agents' own YouTube accounts in an embedded player, and an embed reports nothing back to the embedding site — watch time, completion and drop-off are measured by YouTube, for the channel owner, and `media.meta` carries no statistics on any of the 13 rows. The two ways to fill this are both real work with a real cost: pull public view counts through the YouTube Data API on the existing re-sync sweep (cheap, but it counts views from YouTube's own surfaces — people who never came near this site — so it is not Realty.tube behaviour and must never be labelled as such), or host video ourselves and measure playback directly (accurate, and a different platform). Until one of those is chosen, a video-engagement number here would be a guess with a percent sign.

Updated today · Based on all activity ever recorded · Based on available Realty.tube data

The newest observation behind this reading is from this platform's own recorded activity, today.

search_demand is cumulative — it holds running totals rather than per-period counts — so this reading covers everything ever recorded. It will not fall when interest falls, which is the property to know about it.

Based on available Realty.tube data — activity recorded on this platform and listings published on it. It is not a registry extract, a survey, or a national index, and it describes this platform's market rather than the whole market.

How this is calculated

Every signal is counted in DISTINCT PEOPLE (or the closest honest proxy — a device, a browser session, an account), never in events. Each count is multiplied by how much the action costs the person doing it — a search counts 1, an inquiry 12 — and the results are added. The bands compare that total with the other places published in the same view, so they are relative to what is on screen.

docs/DISCOVERY-SPINE.md §9 — Demand

Select an area

Choose a mark to see its demand broken down by signal — searches, views, shares, saves and inquiries, each with the number of distinct people behind it.

Demand against supply

Not classified in this view. Sorting places into high and low demand against high and low supply needs at least 4 published places to cut between — with fewer, the median falls between two markets and calling one of them “hot” describes the fact that there are two of them, not the market. The demand index and the listing count are both shown above and can be read against each other directly.

Busiest in view

BengaluruMeasured, not ranked
Property views · Saves · Searches — 1 listing on the market here

Every place listed here clears the floor of 5 distinct people on at least one signal. Places below it are drawn on the map in grey and carry no reading. Rows are ordered by index points; the words are the bands those points fall in.