GlobeMetrics
EN

Try “Pune”, “Poona” or “पुणे”

How the numbers are made

Every number on this site has one of four statuses, shown as a chip beside it:

  • Census 2011: the final count by the Office of the Registrar General & Census Commissioner (ORGI).
  • Official projection: the RGI Technical Group on Population Projections (July 2020), population as on 1 March of each year.
  • Estimate: our own figure (method M1), used only where no official figure exists: districts and cities after 2011.
  • Census 2027 provisional: will appear when ORGI publishes.

Precedence: Census 2027 final, then Census 2027 provisional, then the official projection, then M1. We never “improve” an official figure.

Sources

See the full source register. Census tables are downloaded from the ORGI data portal, checksummed (SHA-256) and parsed by scripts that fail the build if any total disagrees with the published one.

Census 2011 on today’s district boundaries

India had 640 districts in 2011 and has 784 today (Local Government Directory, 2026-10-04). Most sites either skip new districts or copy the parent district’s figures. We rebuild them from the smallest census units:

  1. Every one of the 6,40,949 Census 2011 villages is placed in its current district through the census code recorded against it in the Local Government Directory (68.62% of India’s population).
  2. Each of the 8,067 Census 2011 towns (or town parts) is placed by the strongest available link: its 2011 sub-district when all of that sub-district’s villages moved together, its LGD village or urban local body record, the LGD sub-district carrying its 2011 code, or (for wholly urban 2011 districts) the district carrying its 2011 code. A short reviewed list covers towns merged into new corporations (e.g. Kulti, Jamuria and Raniganj into Asansol).
  3. Units that cannot be linked (0.24% of the population, mostly un-coded villages in the North-East) go to the district that received most of their 2011 sub-district.

Every rebuilt state total equals the published 2011 total exactly. A district gets its own page only if no more than 2% of its rebuilt population depends on step 3. Of 784 districts, 429 are unchanged since 2011, 342 are rebuilt, and 715 have pages. Delhi’s districts, redrawn on 1 January 2026 along municipal zones, cannot be re-cut from 2011 data and appear only as table rows.

Older figures (1951–2001) for rebuilt districts apportion each 2011 parent district’s A-2 series by the share of its 2011 population now in the district; they are labelled “reconstructed”. Areas of rebuilt districts apportion 2011 district areas by rural population share and are marked ≈.

Estimation method M1

All estimates are dated 1 March, the census reference date.

India and States

No estimate: the official projection is used as published. States formed or merged after 2011 (Telangana, Andhra Pradesh, Jammu & Kashmir, Ladakh, Dadra & Nagar Haveli and Daman & Diu) use the Technical Group’s own series.

Districts: damped share-trend, constrained to the state

  1. s2001, s2011: the district’s share of its state’s population on 2011-consistent boundaries.
  2. r = clamp( ln(s2011 / s2001) / 10, −0.02, +0.02 ) per year.
  3. ŝ(t) = s2011 × exp(0.5 × r × (t − 2011)). The trend is halved because it is 15 years old.
  4. s(t) = ŝ(t) / Σ ŝ(t) over the state’s districts, so district estimates add up exactly to the official state projection.
  5. P(t) = s(t) × Pstate(t), rounded to the nearest 1,000.

Cities: scaled own trend

  1. gc = ln(P2011 / P2001) / 10 for the city (and separately for its urban agglomeration).
  2. gs,past = ln(State2011 / State2001) / 10; gs,future = ln(Projectiont / State2011) / (t − 2011).
  3. g = clamp( gc × gs,future / gs,past, 0, 0.04 ); P(t) = P2011 × exp(g × (t − 2011)).
  4. If the city’s area changed by more than 15% between 2001 and 2011 (so the two counts are not comparable), the city grows at the rate of the M1 estimate for its district(s).
  5. Rounded to the nearest 10,000 (nearest 1,000 under 1 lakh), always on 2011 city limits.

Estimates are not forecasts of what Census 2027 will find. Their main weakness is migration since 2011, which no published source measures at district level.

Religion, age, languages, housing and income

These come from four Census 2011 tables (C-01 religion, C-14 five-year age groups, C-16 mother tongue, Houselisting HH-14 amenities) and, for state income, the National Statistical Office series compiled in the RBI Handbook of Statistics on Indian States. They describe 2011 (income: up to 2024-25) and are not projected forward.

  • Districts unchanged since 2011 show the official district row as published.
  • Districts redrawn after 2011 are added up from smaller 2011 units placed by the crosswalk: housing from every village and town, weighted by its households (on unchanged districts this reproduces the official figures to within 0.2 percentage points on average); religion and languages from whole 2011 sub-districts; age from the 2011 parent districts in proportion to population. Where a 2011 unit was split between districts, its share is apportioned by population and the figure is labelled “approx.”
  • States use the official state row, except Andhra Pradesh, Telangana, Jammu and Kashmir, Ladakh and Dadra and Nagar Haveli and Daman and Diu, which are added up from their districts. HH-14 has no state rows, so state and India housing shares are the household-weighted average of the official district rows.
  • Languages follow the Census grouping of mother tongues into languages. Median age is interpolated within five-year groups.
  • Religion is shown as an official table only, with no commentary, no rankings language and no advertising next to it.
  • Per capita income is per capita Net State Domestic Product at current prices; real growth uses constant 2011-12 prices. Recent years are advance or quick estimates and are revised by the states.

Checks that block publication

  • Districts add up to their state; states add up to India (Census 2011, exactly).
  • District estimates add up to the official state projection.
  • Recomputed sex ratio, literacy and density agree with the published values.
  • Values stay within sane ranges (sex ratio 600–1,200, literacy 30–100%).
  • Any official figure that changes by more than 2% between builds stops the build until reviewed.
  • Every page must carry at least 12 figures, a table of five or more rows, a chart and three data-based observations, or it is not published.

Census 2027

When ORGI releases results, they replace projections and estimates on the existing pages the same day. We will never publish leaked or unofficial counts. When only state totals are out, district estimates are re-scaled to them and labelled “re-benchmarked to Census 2027 provisional”.

Maps

State outlines follow the official map of India published by the Survey of India, including the whole of Jammu and Kashmir and Ladakh.