Skip to main content
Bihari AddaBihar tech & AI tracker

Methodology

A dashboard that shows its working

A curated public-interest tracker, not a directory or a forecast. It prioritises auditability over volume.

Principles

Evidence before reach

A record needs a dated primary source or a reputable, attributable report before it is published.

Statuses are not outcomes

Announcements, MoUs, pilots, operational systems, and verified outcomes are distinct states.

Unknown stays unknown

Funding, jobs, adoption, and training are left blank when the source does not establish them.

Every record is reviewable

Records keep each source with its publisher and publication date, a link where one is public, and a list of unverified claims.

Update workflow

  1. Collect a primary source or attributable report and record its publication date and, where one exists, its URL.
  2. Extract only claims supported by that source; label targets and planned work as such.
  3. Assign a status, confidence level, and geography.
  4. Review the record before publishing; preserve gaps rather than filling them with assumptions.

Current coverage

Every dataset follows the same rules, and each is deliberately small: it grows through recurring verification, not bulk imports of unsourced listings.

That is 81 records in total, backed by 235 dated sources. Every funding round, recruitment notice and government figure links to one, and each figure says whether it is an official statistic, a government claim or a target.

The downloads above are the same files the dashboard reads, so they cannot drift from what these pages show. Each carries its own as_of date and research principles, and all are free to reuse under CC BY 4.0 with attribution.