The Full Vision — What Our Rivers Will Become

The simple version

India has thousands of rivers.
They are documented in fragments — a Wikipedia page here, a government PDF there, a WRIS layer somewhere else, a news article about a specific flood, a temple website about a specific ghat.

No single place brings it all together.
No single place organises it uniformly.
No single place serves the student, the researcher, the devotee, and the activist with the same information structured in a way that makes sense to each of them.

Our Rivers is being built to be that place.

A clean, uniform, structured encyclopedia of India’s rivers.
Every river with the same pages. Every page with the same fields.
Factual. Data-backed. Consistently maintained.
Sensitive to the cultural and spiritual identity of these rivers.
Accessible to someone with no technical background.
Rigorous enough to be useful to a researcher.

That is the simple version.

The complex version

The data backbone

Every river page in this section is generated from a structured database — not written by hand.

In the first version, this database is a PostgreSQL instance holding a table of rivers with 15 confirmed fields per row.
The Astro/Starlight site queries this database at build time and generates one page per river, automatically, using a shared layout template.

This means:

Every river page has identical structure
A contributor adding a new river only needs to insert a database row — not write a single line of markdown
Fixing an error across all rivers is a single database update, not editing 100 individual files

As scale increases, the database will evolve from a simple Postgres table to a more sophisticated multi-source data system — drawing from official portals, contributor inputs, automated scrapers, and eventually AI/ML pipelines.

The three data layers

Not all data on a river page behaves the same way.
Data is organised into three clearly separated layers:

Layer 1 — Static
Facts that do not change: origin point, length, states covered, zone, river type.
These are entered once, validated, and remain stable.

Layer 2 — Periodic
Data that changes but not in real time: pollution index, water quality category, last known monitoring station readings.
These are updated via an automated weekly build — the site rebuilds from the latest database state and each page shows a “last updated X days ago” badge.

Layer 3 — Live
Real-time readings: dissolved oxygen (DO), biochemical oxygen demand (BOD), turbidity, flow rate where available.
This layer is handled entirely by Triveni-net — a separate, parallel project building an AI/ML-powered live river data service.
On each river page, a React component mounts at the top and pulls live data from Triveni-net independently. The wiki never owns or stores this data.
If Triveni-net is offline or a river has no live sensor, the component handles that state gracefully.

The map

Each river will eventually have a custom interactive map — not a Google Maps embed, not a static image.
A purpose-built flowing diagram that shows:

The river’s course from source to mouth
Its major tributaries in context
Pollution hotspots (fed from Layer 2 data)
Flow density and seasonal variation
Live sensor locations (fed from Triveni-net)

This map component will be shared across the entire Our Rivers section — the same component, configured differently per river via its data.

Roadmap: v1

The map is a future milestone. Current pages reserve space for it with a placeholder.

The automation pipeline

Manual data entry works for 3–4 rivers.
It does not work for 300.

The long-term pipeline is designed in stages:

Manual (now): team enters data directly into the database
Contributors (v0.1): a structured contribution system where verified contributors add river data via a submission form, reviewed before publishing
Semi-automated (v1): scripts pull from official sources (India-WRIS, CPCB, state pollution boards) and pre-fill fields for editorial review
AI/ML pipeline (v2+): automated extraction from satellite data, news, research publications, and sensor networks — human editors validate and approve before any data is published

What makes this different

There are existing resources for Indian river data. None of them do what this does.

Resource	What it does well	What it lacks
India-WRIS	Authoritative official data, rich GIS layers	Designed for technical experts, no narrative, no cultural context
Wikipedia river pages	Good general overview	Inconsistent structure, no live data, no pollution focus
CPCB portals	Official pollution monitoring data	Very technical, no geographic or cultural context
State irrigation department sites	Basin-specific depth	Fragmented, inconsistent across states, rarely updated
News archives	Strong on events and crises	Not structured, not searchable as reference

Our Rivers is not competing with any of these.
It is the layer that sits above them — interpreting, organising, and presenting their data uniformly, with cultural sensitivity and multi-audience accessibility.

Where possible, Our Rivers links back to these sources so that users who want to go deeper can go directly to the authority.

What this section will feel like when complete

A reader arriving at /our-rivers/kaveri/ will see:

A zone overview with aggregate stats for the entire Kaveri basin and a map showing all rivers in the zone
A clean grid of every river in the Kaveri zone, each with a small card: name, type, length, pollution index, and status
Clicking any river opens its dedicated pages — an overview tab with structured facts, a health tab showing the latest pollution data and a live sensor strip from Triveni-net, a hydrology tab describing seasonal and climatic behaviour, and so on

The entire experience will feel like reading a well-maintained, living reference — somewhere between an encyclopedia, a data portal, and a field guide.

That is what Our Rivers is being built to become.

Continue reading: How We Organise Indian Rivers →