Collaboration ask · not a data scrape

Partner GIS ask — MEA / MWA / PSI / GISTDA

Open-data Phase B RF is frozen at v1.56. Catalog grew to 54 BMR / 49 unique; denser-label retrains and spatial-block CV held. Further open churn will not close the Sai Noi / utility-pit FN class without partner layers.

Priority 1

MEA — underground electrical GIS

Sai Noi hard false-negative is cable-pit / utility class. OSM towers ≠ manholes.

  • Manhole / chamber / cable-route inventory (BMR subset OK)
  • Any exportable GeoJSON / GPKG under NDA
  • Coordinate + asset age if available
Priority 1

MWA — pipe / water-main GIS

Many verified voids sit on drainage / water-main corridors.

  • Pipe age / material / diameter layers
  • Leak / repair history if shareable
  • Alignment with BMA drainage where joint
Priority 2

Numeric InSAR PSI

LiCSBAS or equivalent GeoTIFF — not screenshots.

  • BMR velocity / coherence grids
  • Documented epoch + CRS
  • Joinable to 200 m research cells
Priority 2

GISTDA / agency flood & RS

Flood extent already optional on dashboard MapServer; deeper products welcome.

  • Validated flood / soil-moisture products
  • Co-labeled events for spatial-block CV
  • Pilot district boundary for desk trial

What partners get back

  • Honest freeze status + Contabo live meta JSON (operational_use=false)
  • Verified news catalog + candidate label queue (human verify before any promote)
  • No auto-CAP; CAP drafts remain human coordination only
  • Science gate: denser labels promote only if dual AP/F1 gates pass vs v1.56

TH · สรุปคำขอข้อมูล

ต้องการชั้น GIS ใต้ดิน (MEA/MWA) และ PSI เชิงตัวเลข เพื่อลด FN กลุ่มหลุมสายใต้ดิน — ไม่ใช่การขอให้เปิด RF เป็นระบบเตือนภัยจริง และไม่มีการออก CAP อัตโนมัติ

GISTDA partner brief RF freeze + FN note Label candidates What we do NOT claim

Refs: docs/117 spatial CV · docs/106 label harvest · gap to ~80 labels needs partner GIS, not more WorldPop churn.