Bahubhashi India

Visualizing India’s Spoken Language Landscape

A spinout from SPIN Lab, Mehta Family School of Data Science and Artificial Intelligence, IIT Guwahati, India

We build AI that listens, from speech, music and breath sounds to street noise, and turns everyday sensing into open data, visualizations and AI-based tools. Students, researchers and industry partners: come work with us

India speaks in many voices. Yet most maps paint each state in a single language and hide how many languages live side by side. In the 2011 Census, people wrote down their mother tongue in their own words:

  1. 19,569names written for a mother tongue
  2. 1,369mother tongues, after linguists sorted them
  3. 121languages they are grouped under, which the maps show

22 scheduled

Listed in the Eighth Schedule of the Constitution. The mother tongue of 96.7% of Indians.

99 non-scheduled

No constitutional status, but at least 10,000 speakers each. The mother tongue of 3.1% of Indians. The remaining 0.15% reported smaller mother tongues that the census did not group under any of the 121.

The grouping of languages follows the 2011 Census document.

We set out to show how linguistically diverse each state is, and to let you hear a bit of it. Speaker counts come from the Census of India 2011 (Table C-16), the voices from Vaani by IISc Bangalore and ARTPARK, and the scripts from Unicode CLDR.

How we made it

We read Table C-16 straight from the census report: the number of people in every state and union territory whose mother tongue is each of the 121 languages, checked against India's 2011 population of 1,21,08,54,977. For each state we measure diversity two ways: the chance that two people picked at random have different mother tongues, and the effective number of languages, how many equally large languages would give the same mix. States are drawn as they were in 2011, so Telangana is part of Andhra Pradesh and Ladakh part of Jammu & Kashmir.

For the voices, we took each state's largest language and picked people from that state speaking it, from Vaani's recordings of people describing everyday pictures. Each clip is trimmed to 5–10 seconds, matched in loudness and listened to before it goes on the map. Where Vaani has no one speaking a language, or no one from a place, the state's panel says what you hear instead. Scripts are each language's usual script according to Unicode CLDR, drawn in Noto fonts.

Bahubhashi India

Language diversity by state · Census 2011

a spinout from SPIN Lab MFSDSAI · IIT Guwahati

Data, credits & disclaimer

Languages
Census of India 2011, Language: India, States and Union Territories (Paper 1 of 2018), Table C-16 and Statement 9. Office of the Registrar General & Census Commissioner, India. censusindia.gov.in
Speech
Vaani by IISc Bangalore and ARTPARK, licensed CC BY 4.0. We trimmed the clips to 5–10 seconds, matched their loudness, and kept only clips we listened to and checked. Languages are as the speakers stated them.
Boundaries
INDIAN-SHAPEFILES (MIT licence), merged to the 2011 census units.
Scripts
Likely script of each language from Unicode CLDR (Unicode License v3). Letters drawn in Noto fonts (SIL Open Font License).
Software
MapLibre GL JS (BSD-3-Clause) and H3 (Apache 2.0).

Disclaimer

© 2026 Neeraj Sharma, SPIN Lab, Mehta Family School of Data Science and Artificial Intelligence, IIT Guwahati, India. The code is released under the MIT licence; the maps, visualizations and text under CC BY 4.0 (details). Data, recordings, fonts and libraries remain under their own sources' terms, listed above.

How the measures are calculated