← All knowledge

Method

Where the data comes from

The full chain of custody from a KSDMA PDF to a number on screen, and what JalaNet refuses to publish.

Reservoirs · 5 min read

The pipeline

Every hour JalaNet looks for the day's dam report PDFs published by the Kerala State Disaster Management Authority — one filed by KSEB, one by the Irrigation Department. Each PDF is downloaded, its text extracted, and each dam row parsed into structured fields. Weather is fetched separately every 15 minutes from an open forecast model at each dam's coordinates.

Every card carries a provenance panel showing the source feed, when the report was filed, when JalaNet ingested it, when weather and forecast were observed, and whether any unit normalisation was applied.

What is never invented

If a value is missing from the source, JalaNet shows a dash. It does not interpolate storage, does not estimate live electricity generation, and does not model statewide consumption. Where an official feed does not exist, the interface says 'not published' and explains why.

Continue reading

  • How a reservoir worksStorage levels, FRL, MWL, dead storage and why a dam at 100% is not automatically dangerous.
  • Reading rainfall forecastsWhat 24 h and 72 h catchment totals mean, why forecast rain matters more than fallen rain, and forecast uncertainty.
  • Inflow, outflow and spillingCumecs explained, how shutter openings are reported, and how to read a rising or falling reservoir.