Public data only
Everything on this map is public. Here is what it cannot see.
The twin of Muang Thong Thani runs on open building footprints, public camera views, agency feeds and the venue's own published calendar. Nothing proprietary, nothing from inside the estate. That is the point: it shows what anyone could build today, and it shows exactly where the answer stops and a measurement is needed.
229,377Buildings drawn
11,436 from OpenStreetMap, 217,941 from Overture Maps (release 2026-09-23.1).
142Public camera views
56 on the Chaeng Watthana corridor, 86 on regional waterways. Image only; no counts.
0Measured sensors on the estate
Parking fill, internal road load and crowd size are modelled. No public feed measures them.
Ten public layers, each with a licence and a limit
Taken from the twin's own source catalogue (data/source-catalog.json, checked 2026-10-15). The third column is the limit the dashboard already prints beside each number.
Building footprints
- Source · licence
- OpenStreetMap, Overture Maps · ODbL 1.0
- Limit
- Heights estimated. 7.7% of added footprints flagged for overlap review. OSM layer is an archive, not live.
Parking areas
- Source · licence
- OpenStreetMap · ODbL 1.0; IMPACT published page
- Limit
- 150 polygons. Capacity and access missing for many. OSM and venue capacities disagree in places; both shown.
Public camera views
- Source · licence
- Municipal and agency feeds via MAHOLAN · public viewing
- Limit
- Redistribution terms are not documented. Coordinates are publisher-reported. A view is not a count.
Water and air
- Source · licence
- HII, PCD via FloodDash and AirDash · public agency
- Limit
- River level is not water depth at a building. Air comes from named stations, not from the estate.
Flood zones
- Source · licence
- GISTDA GFlood; Nonthaburi catalogue (data.go.th)
- Limit
- Risk inventory is from 2024. The catalogue states no licence, so nothing new is redistributed.
Traffic and events
- Source · licence
- Longdo / iTIC incidents; IMPACT public calendar
- Limit
- No measured speeds, no attendance. Road load inside the estate is a model, stored every 15 minutes for history.
Aerial context
- Source · licence
- EOX / Copernicus Sentinel-2 cloudless, 2016
- Limit
- About 10 m pixels. Cannot show a parking bay. Not used to draw lots.
Nine things nobody can currently see
Each row names a decision that is made blind today, the best public stand-in, and the smallest measurement that replaces the stand-in. Cost is a class, not a quote: low means off-the-shelf units, high means a procurement.
Parking fill on event days
- Today
- Modelled from hall attendance: 62% by car, against about 12,500 published spaces. Nothing measured.
- Measure with
- Entry and exit counters at the main lots (P1–P5 first). Cost: low–mid
- Unlocks
- Real overflow warnings, shuttle timing, and a number to test the model against.
Roads inside the estate
- Today
- Private roads, so no provider has data. 7 modelled segments. Only Chaeng Watthana has real incident feeds.
- Measure with
- Radar or probe counters at the gates and two internal junctions. Cost: low–mid
- Unlocks
- Measured queue length at event exit, which the model can only guess.
Crowd and attendance
- Today
- The calendar says what is on, not how many came. Shown as illustrative.
- Measure with
- Gate counts from ticketing, and MRT Pink Line ridership at MT01 and MT02. Agreements, not hardware. Cost: low
- Unlocks
- Every other model in the twin starts from this number.
Noise
- Today
- None continuous. The data bible counts 490+ events a year at 90–110 dBA at source.
- Measure with
- Three to five noise monitors on the residential and lake edges. Cost: low
- Unlocks
- Evidence for complaints and for event permits, in both directions.
Air at the estate
- Today
- One citizen station (WAQI) plus modelled PM10 and NO₂ from CAMS.
- Measure with
- Three to five low-cost PM2.5 units, calibrated against the one reference station. Cost: low
- Unlocks
- Street-level air, not district-level.
Khlong and flood depth
- Today
- River gauges and satellite flood polygons. The khlong's last water-quality sample is reported as 13 years old.
- Measure with
- Ultrasonic level sensors at low points, one rain gauge, and the 86 waterway cameras already public. Cost: low
- Unlocks
- Depth where it floods, not a gauge kilometres away.
Camera counting
- Today
- 142 views, no counts. The image check runs on public frames; 8 cameras still fail it.
- Measure with
- Counting analytics on a few owned cameras at gates and the lake path. Cost: mid
- Unlocks
- People and vehicles counted, with no face data kept.
Emergency response
- Today
- Pak Kret fire and rescue publish nothing. The data bible notes 46 officers per 100,000 residents.
- Measure with
- A data-sharing agreement for incident times and locations. Cost: low
- Unlocks
- Response time to the estate on event days.
Complaints
- Today
- Traffy Fondue holds about 865,000 reports. No public per-area view.
- Measure with
- An area export for the estate boundary. Cost: low
- Unlocks
- What residents report, mapped next to what the twin models.
Start with agreements. Add hardware only where a model fails.
Each step shows up in the twin the day it exists, because the model and the measurement share one screen.
Step 1 · no hardware
Agreements
- Gate counts from ticketing
- MRT ridership at MT01 and MT02
- Fire and rescue incident log
- Traffy Fondue area export
Step 2 · low cost
Small sensors
- Air units, noise monitors
- Water-level sensors and a rain gauge
- Counters at the main lot entrances
Step 3 · mid cost
Counting cameras
- Analytics at gates and the lake path
- Replace the road model with measured load
- Alerts tied to a named owner
Limits, stated plainly
- No prices are quoted. Cost classes are a guide for sizing the conversation; a supplier quote replaces them.
- The camera feeds are public to view. Redistribution terms are not documented, so the twin links and checks them but does not claim a licence.
- The flood-risk catalogue states no licence. Only previously documented references are kept.
- The data bible lists 12,684 parking spaces; the venue page and model use 12,561 across 11 facilities. The two have not been reconciled.
- Building coverage is a mapped study area, not a verified municipal census. Heights are estimates.
- Modelled values are labelled as modelled in the twin. A measurement, once installed, replaces the label, not the model's history.