How People & Consent’s figures were produced

The instrument takes one proposal from India’s environmental clearance register and assembles the public record about the place where it would be built. It asks who lives there and what the law requires of a project in that place, a question about one site that needs no comparison between districts, so the instrument gives no score or ranking and applies no weights.

It holds 93,000 proposals, at every stage from filing to decision. A further 26,140 records are left out because the portal has delisted them. They carry no decision and appear in no count on this site.

Every population figure on a profile depends on where the project is placed, so this section sets out the method in detail.

The finest location the register’s search API returns is a project’s state. It has no district field, so the district is read from the application form each proponent files.

That form, the Common Application Form, records the location the proponent declared, with district, sub-district, village and plot numbers. Reading it directly places 99.7% of proposals in a district, and 94.96% in a sub-district as well. The confidence given to each route in the table reflects how specific the statement it rests on is: a named village or sub-district is the most specific, a district named alone is less so, and a district inferred from the project name is the least.

Two features of the form produce answers that look plausible and are wrong, which makes them harder to catch than an outright failure.

The wrong district field. A Common Application Form carries three districts: the project’s, the proponent’s registered office, and the applicant’s address. The last two are often a large city hundreds of kilometres from the site, so only the project’s district is used here.

Two code systems in one record. Within a single filing the village and sub-district codes are Census 2011 identifiers while the district code is a Local Government Directory code. The two numbering systems differ, and treating them as one would join every record to the wrong place. The village code is preferred throughout because it is a Census identifier, and so it resolves correctly even for districts created after 2011. Morbi, created in 2013, resolves through its Morvi sub-district to its 2011 parent, Rajkot, with no table of district changes needed.

1,815 proposals name sites in more than one district, and most of them are linear projects such as highways, transmission lines and pipelines. Their profile describes the primary district, the one named by the most plots unless the reconciliation described next moves it, and states that the project spans other districts too.

Three statements of place, reconciled. A filing states where a project is in three ways: the place codes on its plots, the district it declares and the coordinates of its site. The codes are preferred because they are the most specific, but a code can be wrong in a way the other two reveal. Where the declared district and the coordinates agree with each other against the codes, they decide, and 128 proposals are placed that way. A town that spans more than one district, such as Greater Mumbai or Greater Hyderabad, carries one code for all of them, so the code alone cannot say which district a site is in. Where such a code was the only basis for a placement, the coordinates or the declared district decide (2,135 proposals), and where neither can, the placement is published as provisional. Where no two of the three statements agree (24 proposals), the proposal stays where its codes put it and its page describes the location as contested. Each profile states which of these routes it rests on.

Socio-economic figures come from the Census of India 2011, obtained through SHRUG, and are added up from sub-district rows to the level being shown. The instrument profiles 640 districts and 5,899 sub-districts.

Census 2011 is 15 years old, but it remains the most recent census for which complete sub-district figures exist. Census 2027 is under way, with its population count scheduled for February 2027, and none of its figures are published yet. Each figure on the profiles states the year of its source beside it.

Each measure’s definition, the formula that produces it, its source and the year of that source are kept in one registry, and the district files carry only the values. The interface joins the two when it draws a figure, which is how every number on the instrument can show where it came from.

The same design guards against one kind of error. A benchmark and the value it is compared with must resolve to the same registry entry, so a rate is only ever compared with a benchmark computed over the same denominator. Where that check fails, the comparison is left out.

Several fields in the source look usable and turn out to be unreliable. They are left out, and the reason for each is recorded here.

• Primary Health Centre counts. The Village Directory column sums to roughly 58,800 nationally against about 24,000 actual centres in 2011, with 30,000 in a single sub-district, which shows the column is corrupt at this level of aggregation. The doctor vacancy rate, a ratio of two columns from the same family, is used instead because the ratio holds up whatever unit the counts were recorded in.
• Electricity, water and bus coverage. At sub-district level these are yes-or-no flags recording whether any village has the service, and 5,590 of 5,834 sub-districts read 1. With 96% answering yes nationally, the flag says nothing about how many people are covered, so it is left off the profiles.
• Counts that are zero across an entire state. Every one of Gujarat’s 26 districts reports zero government middle schools while Jamnagar alone reports 1,217 primary schools. Gujarat’s primary schools run to standard 8, so the state has no separate middle schools to count. Where a count is zero across a whole state but substantial nationally, the value is withheld, because a zero would tell readers that the schools do not exist.
• Doctor vacancy where the source contradicts itself. In some districts the Village Directory records more doctors in post than posts sanctioned, which yields impossible rates such as −466%. The rate is withheld in those districts and both raw counts remain visible.

Each entry on statutes and standards was transcribed by hand from the provision it cites and checked against the primary source, and each stands independently of the population data. The reference holds 15 obligations, all 15 of them verified against a primary source, and was last reviewed on 2026-09-25.

Whether a place lies in a Scheduled Area under the Fifth Schedule is decided by Presidential Order under paragraph 6 of that Schedule. No authoritative list exists in machine-readable form. When checked in August 2026, the Ministry of Tribal Affairs page that once carried one returned an error (HTTP 404), and the Ministry of Panchayati Raj’s state-wise page gives no Order citations and leaves its coverage columns empty. The Orders themselves were therefore read.

639 of the 640 districts have a determined status, reached in one of four ways. A district an Order names takes its status from that Order. Where an Order covers the State but omits the district, the omission settles it, because the Order lists every Scheduled Area in its State. Where no Presidential Order reaches the State, the district lies outside any Scheduled Area. In Assam, Meghalaya, Tripura and Mizoram, Article 244(1) settles the matter, because it applies the Fifth Schedule to Scheduled Areas in every State other than those four. The remaining 1 district is published as not determined, for the reason given below.

Districts drawn after an Order. Every Order defines its area by the territorial divisions that existed on its date, and drawing a new district later leaves the scheduled ground unchanged. A later district’s status therefore follows from where its ground lay on that date. A district carved wholly from one the Order names in whole is wholly scheduled, as are Rayagada, Malkangiri and Nabarangapur, carved from Koraput, and Bijapur, Narayanpur, Alirajpur and Khunti. A district that takes in part of an area the Order names is partly scheduled, and one that takes in none of it is outside. Each such result is checked against the Ministry of Panchayati Raj’s list of notified Fifth Schedule Areas, which names every district decided to be inside and none decided to be outside.

Andhra Pradesh and Telangana are governed by the 1950 Orders C.O. 9 and C.O. 26, which schedule the agency tracts of the old Madras districts and villages of the old Hyderabad State, district by district. The districts they reach are partly scheduled: East Godavari, West Godavari, Visakhapatnam, Srikakulam and Vizianagaram in Andhra Pradesh, and Adilabad, Warangal, Khammam and Mahbubnagar in Telangana. The rest of both States is outside. The text of C.O. 9, the Andhra Pradesh High Court’s list of the agency districts and the Ministry of Panchayati Raj’s list all name the same set.

One district remains undetermined because two official sources conflict over it: the Order does not name it, while the Ministry of Panchayati Raj lists it. The note on each such district’s page cites both. In a partly scheduled district, a site is placed inside the Scheduled Area when its filing puts it in a sub-division the Order names in whole. In every other case the district’s status cannot settle the question, and the page says so.

Two claims that circulate widely are corrected here. C.O. 192 of 2003 is titled for Chhattisgarh, Jharkhand and Madhya Pradesh, and its schedule lists thirty areas in Chhattisgarh and Madhya Pradesh, while Jharkhand’s areas are scheduled by C.O. 229 of 2007. Odisha lies outside C.O. 192 and remains under C.O. 109 of 1977. Rajasthan’s C.O. 114 of 1981 was expressly rescinded by C.O. 270 of 2018, although the Ministry of Tribal Affairs’ own compendium, compiled before 2018, still lists the 1981 Order as operative.

Sixth Schedule areas in Assam, Meghalaya, Tripura and Mizoram fall under a separate constitutional regime, administered through Autonomous District Councils, and the Panchayats (Extension to Scheduled Areas) Act, 1996 (PESA) does not apply to them. They carry their own label, distinct from both scheduled and unscheduled.

The location is checked against a second document. Everything on this instrument depends on knowing where each project is, and that knowledge comes from a single source, the form the proponent filled in. The placement is therefore tested against the regulator’s signed clearance letter, a separate document written by a different party, which describes the site in prose.

On a random sample of 150 proposals, the district this instrument reports matched the district named in the letter in 86.15% of the 65 cases where the letter could be read and named one. Of the remaining 85, 11 name no district this check could match, and 74 could not be retrieved or read. Those failures reflect the limits of the check itself, and they say nothing about the letters.

Only the text near a location cue in the letter is read. A letterhead carries the issuing authority’s own address, such as Gandhinagar, Bangalore or Chennai, and reading the whole letter would mistake that city for the site.

The site coordinates check every location. Each filing’s mapped blocks are tested against the Census 2011 district boundaries: 89,313 fall inside a district the proposal is placed in and 1,910 on its edge; 928 fall elsewhere, and each of those pages says where the coordinates fall and which statements of place the profile follows instead.

A rebuild that fails its checks publishes nothing. Every measure declares the range it can plausibly fall in, and a value outside that range stops the rebuild. Figures that must add up are checked, so the cultivator and agricultural labourer shares have to sum to the land-dependent share. A sample of the published figures is also worked out a second time from the original sub-district records by a separate calculation, which catches a mistake in the first.

The most recent rebuild recorded 0 failed checks and 0 figures the second calculation could not reproduce. As a test of the whole chain, national literacy worked out from these figures is 72.98% against the Census of India’s published 72.98%. The match is consistent with a literacy denominator of the population aged 7 and above, which is the Census definition; computed over the total population, the rate would be materially lower.

The time a proposal took to be decided is measured from the first step in its appraisal history, the dated record of steps the portal keeps for each proposal, to the step at which the authority granted or refused what was applied for. The register’s own submission date is published beside it but never measured from, because it moves forward whenever a proposal is submitted again. Days on which the portal had a proposal automatically delisted are subtracted and published separately, so the first filing date plus the days to decision plus the days delisted equals the decision date on every decided proposal.

The source timestamps are in UTC and are converted to India Standard Time before a date is taken. The change log dates each event by the same history: a filing or a decision by the end of its step, and a query, closure or other step by its start. Each event also records the build in which it first appeared, so a reader following the log can ask for everything that is new since a previous visit, including steps the portal recorded late.

• 283 proposals could not be placed in any district. For 1 of them the Ministry’s portal serves no Common Application Form at all, and 0 have not been read yet. The rest were read and record no village, sub-district or district, and their project names contain no district that matches the Census gazetteer. They are reported as undetermined, and their profiles show no state averages in place of district figures.
• The record of whether the Forest Rights Act was complied with before forest land was diverted lies outside the public data. That certificate is part of the forest clearance application completed by the District Collector, and the portal serves that part only to signed-in users. The instrument shows the public part instead: the forest proposal the environmental form names and its current stage in the forest register.
• Counts of affected and displaced families are the figures each proponent declared in its application. They are the proponent’s own statements, and the pages label them as declared wherever they appear.
• Whether a place lies within a Scheduled Area under the Fifth Schedule is a legal determination made by Presidential Order, and no authoritative list exists in machine-readable form. Where the transcribed Orders do not settle a place, the instrument reports its status as not determined. It never infers the status from the share of Scheduled Tribes in the population, because a Scheduled Area exists only where an Order creates one, and a place’s tribal share does not decide whether it does.
• The register as this instrument reads it begins with proposals submitted in September 2022, so it covers only the recent part of the history of environmental clearance in India.
• Population figures are shown for the sub-district or district even where a village is named. The village identifies the place, and the profile then reports the record for the sub-district or district, the level at which the Census measures used here are published.

• PARIVESH environmental clearance register
  Ministry of Environment, Forest and Climate Change
  Updating source · read 28 October 2026
  proposals submitted 2022 onward
• PARIVESH Common Application Form
  Ministry of Environment, Forest and Climate Change
  Updating source · read 28 October 2026
  project location as filed by the proponent: district, sub-district and village, plus declared land area, affected and displaced families
• Census 2011 Primary Census Abstract
  Office of the Registrar General of India, via SHRUG
  Collected 2011 · version 5.0
• Census 2011 Village Directory
  Office of the Registrar General of India, via SHRUG
  Collected 2011 · version 5.0
• Socio Economic and Caste Census 2011, rural
  Ministry of Rural Development, via SHRUG
  Collected 2011 · version 5.0
  rural households only: land ownership, income band, main source of income and salaried employment
• PARIVESH appraisal history
  Ministry of Environment, Forest and Climate Change
  Updating source · read 28 October 2026
  every dated step a proposal has passed through, from submission to decision. Individual officers are named in the source and are not recorded here.
• Census 2011 district boundaries
  datameet
  Collected 2011
  district outlines as they stood at the 2011 Census, used to test filed coordinates against the district a proposal names

Everything here is produced by one process that can be run again from the same sources. Each run records the match rate for every placement route it used and every check that failed.