How a state report is built
Every report on this site is the same shape: the same chapters, in the same order, drawn from named sources and stamped with the month it was compiled. That uniformity is the point — it is what lets two states be read against each other instead of merely read. This page documents the method, and says plainly where it is thin.
- 9/36
- Reports published
- 17
- Chapters per report
- 25
- Organisations cited
- 297
- District profiles
The template
One structure, every state
All 9 published reports carry exactly 17 chapters, and a given chapter number means the same thing in each of them — chapter 9 is the economy whether you are reading Kerala or Manipur. A handful of chapter titles name their own state, which is why they appear below as <State>.
- 01At a Glancewording varies
- 02History
- 03Government, Politics & Administration
- 04Geography & Districts
- 05Population & Demographics
- 06Culture, Festivals & Cuisine
- 07Education & Research Institutions
- 08Cities & Urban Areas
- 09The Economy
- 10Industries & What Drives Them
- 11Transport & Connectivity
- 12Environment, Wildlife & Biodiversity
- 13Problems Facing <State>
- 14Solutions & Policy Responses
- 15What <State> Can Learn From Other States & Regionswording varies
- 16Future Growth & Potential
- 17Notable Facts
What a chapter is made of
A small, fixed block vocabulary
Chapters are not free-form documents. Each is a list of typed blocks, and only 7 types exist: figure, heading, note, paragraph, quote, statgrid, table. Every block is validated against a schema when the site loads its data — a malformed or mistyped block fails the build rather than rendering blank three states later. That constraint is what makes a report machine-readable: the compare view, the district pages, the timeline and the quiz all read the same blocks rather than re-parsing prose.
Sourcing
Cited at report level, and at the quote
Each report names the organisations it draws on — 25 distinct ones across the site — and each quoted passage carries its own attribution, 18of them so far. Those are the two granularities that exist. An individual statistic in a stat grid is covered by its report's source list rather than by a footnote of its own, so tracing one number means checking the named source for that figure. The full index is on the sources page.
Dating
What “compiled” does and does not mean
Every report is stamped with the month its numbers were gathered — currently August 2026 and July 2026. That is the date of the compilation, not of the data. Census figures are 2011 regardless of when they were copied; NFHS-5 is 2019–21; GSDP series lag by a year or more. Where a figure's own vintage matters, the report states it next to the number instead of relying on the compilation stamp. Each report also links the source document it was compiled from.
Districts
One shared boundary topology
District boundaries come from the Local Government Directory release rather than from any single report's own map, and are stored as one shared topology: two neighbouring districts reference the same edge instead of each carrying an independent outline. Independent outlines were the earlier approach and they left visible seams between districts, because two separately-simplified copies of the same border do not agree. The boundary release is itself dated — districts created after it fall back to their parent district rather than being invented. The geometry covers 785 districts nationwide; 297 of them have a written profile attached.
Limits
Where this is weakest
- Coverage. 9 of 36 states and union territories are published. Any national picture drawn from this site is a picture of those 9.
- Citation granularity. Attribution is per report and per quote, not per statistic. A reader checking one figure has to identify which of the named sources carries it.
- Mixed vintages. A single chapter can set a 2011 census count beside a 2024 budget line. Both are dated in place, but the comparison between them is not like-for-like.