UDISE+ — Unified District Information System for Education (India) Data

It is India's count of its own schools, collected because the Ministry of Education needs it to run the system - which is why the universe is effectively complete rather than a volunteer panel. Roughly 1.47 million schools report each cycle - 1,466,682 for 2025-26 - employing 10,273,020 teachers and enrolling 247,219,766 students from the foundational stage through secondary.

Each school answers an eleven-section Data Capture Form: profile and facilities, then student and teacher detail, entered online during the academic year and validated by MIS officials at block, district and state level before national consolidation. The design decision that matters downstream: the school is the unit of data collection, but the district is the unit of data distribution - so the finest grain public reporting guarantees is district and block, with school-level visibility reserved for the report-card and micro-data surfaces.

Around the core sit four public shelves. Know Your School publishes per-school report cards. The Dashboard & Reports application, aligned to NEP-2020, holds 86 statistical reports across school, teacher, student and infrastructure domains. An archive dashboard preserves 2012-13 through 2021-22. Seven micro-data modules round it out. Which shelf holds your question is exactly the kind of thing a sample settles before anything recurring starts.

Captured UDISE+ records for the 2025-26 cycle, exactly as they arrive:

# edu-highlights summary row - one row per geography per academic year
# captured values, shown to fix the shape

regionName   India            yearId     12            (2025-26)
totSch       1466682          ttch       10273020
tenrF2Sec    247219766        pgroundPer 81.9
elecPer      94.99            waterPer   99.49         compPer 69.94

# report R81 - Number of Schools by School Management and School Category
# one row per management type x location, eleven category columns

report_code  R81              sch_mgmt_name  Department of Education
locn_name    All India
cat1 465992   cat2 152039   cat3 16929    cat4 64398    cat5 21673
cat6 16413    cat7 15723    cat8 11750    cat9 0        cat10 11812   cat11 2336
total 779065

# report R101 extract - Gross Enrolment Ratio cells by level x gender x social category

report_code           R101
indicator             Gross Enrolment Ratio (GER)
ger_primary_all       103.39      ger_elementary_all   100.13
ger_secondary_all_sc  84.91       ger_upper_primary_all (null - not reported)

The first block is the highlights row - one line per geography per year carrying the headline denominators. The second and third are report-table rows: R81 counting schools by management type across eleven category columns, R101 giving Gross Enrolment Ratio cells by level and social group. Note the null - upper-primary GER simply is not populated in that extract, and knowing which cells go empty is part of knowing the data.

A UDISE+ record divides into identity, classification and measurement, as tabled below. Four earn special mention.

udise_code is the entire integration strategy. Schools open, close, merge and rename; the code persists through the School Directory. Any analysis joining infrastructure to enrolment to teachers starts here, and any external dataset worth joining to Indian education keys to it too.

sch_mgmt_name encodes who runs the school - Department of Education, Tribal Welfare and the rest. In a system where government, aided and private-unaided schools behave very differently, it is the filter column behind most headlines about Indian schooling.

elecPer / waterPer / toiletPer come paired with functional variants suffixed Fun. A school can report a facility that does not work; the pair separates infrastructure existence from usable infrastructure, and collapsing them quietly inflates readiness.

ger_primary_all is one cell of a 48-cell grid - level by gender by social category in report R101. Quoting GER without naming the cell is how enrolment debates talk past each other.

• Geography: all States and Union Territories, resolving to district and block in public reporting - the system's own distribution grain - with school-level visibility through the report-card and micro-data surfaces. State comparisons work mechanically; anything below block requires naming your schools.
• Temporal: the current cycle is Academic Year 2025-26, and the archive preserves 2012-13 through 2021-22, with open year-master lists spanning 2018-19 to 2025-26. Depth is genuine but uneven - earlier cycles carry narrower variable sets than recent ones, and the NEP-2020 stage structure re-cut the class grid mid-series, so a long panel is assembled with attention to what each era can answer.
• Granularity: one DCF per recognized school per year across eleven sections, aggregated into 86 published statistical reports at national, state, district and block levels. School-level collection feeding district-level publication is the defining constraint: analyses that need named schools use different surfaces than analyses content with districts.

API, files, or your warehouse. Daily, weekly, or hourly.

Government MIS output rewards nobody who meets it raw. Report families land shaped for administration - codes for labels, percentage columns beside their functional twins, stage boundaries that moved when the NEP framework arrived, and archive years whose variables differ from current ones. We map the field dictionaries once, decode the enumerations, hold udise_code continuity across the series and hand you typed tables that load without a cleanup pass. You pick the channel and the cadence; the definitions travel unchanged through all three.

• Edtech and school-services market sizing - school and enrolment counts by district and management type make defensible TAM models for anything sold to Indian schools (market sizing playbook).
• Citation-grade research and policy work - figures citable to the Ministry of Education's own statistics system clear peer review (citation-grade research).
• Infrastructure and CSR targeting - facility gaps at district and block level direct capital and programs to measurable need.
• Teacher-supply analysis - staffing by management type, qualification and stage flags shortage and surplus districts ahead of hiring cycles.
• Model features - a decade-plus of school-year observations trains enrolment, dropout and infrastructure-gap models (ML model training).
• Investment due diligence - chain-scale claims checked against census counts, enrolment and staffing before term sheets move.

Market researchers and consultants get the census denominators behind every India school-market chart, citable to the district (market researchers page). Investors and quant researchers get sector fundamentals - enrolment trends, staffing ratios, infrastructure levels - on one stable school key (investors and quants). Competitive intelligence teams watch enrolment shift across named districts and management types (competitive intelligence page). Data scientists inherit a keyed school-year panel that joins without fuzzy matching (data scientists page). Journalists and academics cite the same system the ministry itself reports from (journalists and academics page). Sales and growth teams plan territories against every recognized school in the country, filtered to the ones that buy (sales and growth teams page).

Vintage note - the archive reaches to 2012-13 but variable depth grows toward the present, and the NEP-2020 stage re-cut sits inside the series. Treat the usable span per section as something to pin in the sample rather than assume from the archive's overall reach.

Functionality note - facility percentages ship beside their functional variants, and we keep both, because a school with a toilet and a school with a working toilet are different facts about readiness.

Grain note - district is the distribution unit by design. Neighboring records below compose with this one instead of overlapping it: global compilations benchmark India against other systems, and commercial directories add the firm-level view the census deliberately lacks.