This dataset is released under the Creative Commons Attribution-NonCommercial-ShareAlike 4.0 licence (CC BY-NC-SA 4.0): research, academic, personal and other non-commercial use is permitted with attribution; adaptations must be shared under the same licence; commercial use is not permitted. The original public source data remains under the Government Open Data License - India (GODL-India). Commercial use: contact surya@judicialmind.ai or hello@judicialmind.ai.

Bharat Health Service Atlas

Licence: CC BY-NC-SA 4.0 — research and non-commercial use only; derivatives must be shared under the same licence. Commercial use: contact surya@judicialmind.ai or hello@judicialmind.ai

Access is approved automatically when the email you enter matches your Hugging Face account email and belongs to the organisation you name; other requests are reviewed manually.

How well each district delivers basic health services, year by year. The atlas covers district × financial year service-delivery indicators from India's Health Management Information System (HMIS), FY2009-10 to FY2021-22, and a transparent composite index, the Health Service Delivery Index (HSDI). Alongside them are Rural Health Statistics (RHS) facility counts, FY2016-17 to FY2021-22, set against population norms.

district_year: 6,306 district-years, 730 LGD districts. Rows per FY: 2009: 26, 2010: 108, 2011: 230, 2012: 399, 2013: 555, 2014: 592, 2015: 592, 2016: 618, 2017: 582, 2018: 579, 2019: 595, 2020: 727, 2021: 703
hsdi_index: 5,637 scored district-years (districts with ≥ 9 reported months). Scored per FY: 2009: 5, 2010: 37, 2011: 195, 2012: 215, 2013: 444, 2014: 569, 2015: 562, 2016: 582, 2017: 545, 2018: 553, 2019: 548, 2020: 684, 2021: 698
rhs_infrastructure: 3,874 district-years over 734 districts. In the latest year (NDAP label "Financial Year 2021"), 203 districts fall short of the sub-centre norm, 173 of the PHC norm and 257 of the CHC norm.

Quickstart

import pandas as pd

# after `huggingface-cli download judicialmind/Bharat-Health-Service-Atlas --repo-type dataset --local-dir Bharat-Health-Service-Atlas`
hsdi = pd.read_parquet("Bharat-Health-Service-Atlas/data/hsdi_index.parquet")   # or "hf://datasets/judicialmind/Bharat-Health-Service-Atlas/data/hsdi_index.parquet"
top = hsdi[hsdi.fy == 2021].sort_values("hsdi", ascending=False)
print(top[["district_name", "state_name", "hsdi", "hsdi_quintile"]].head(5).to_string(index=False))

rhs = pd.read_parquet("Bharat-Health-Service-Atlas/data/rhs_infrastructure.parquet")
latest = rhs[rhs.fy == rhs.fy.max()]
print("CHC shortfall:", int((latest.chcs_shortfall > 0).sum()), "of", len(latest), "districts")

 district_name state_name      hsdi  hsdi_quintile
     Cuddalore Tamil Nadu 79.597991            5.0
       Kaithal    Haryana 79.540602            5.0
    Tirupathur Tamil Nadu 79.282352            5.0
      Ariyalur Tamil Nadu 79.196269            5.0
Mayiladuthurai Tamil Nadu 79.052797            5.0
CHC shortfall: 257 of 719 districts

With DuckDB:

import duckdb

print(duckdb.sql("""
    SELECT fy, round(median(institutional_delivery_share), 3) AS inst_delivery,
           round(median(c_section_rate), 3) AS c_section, count(*) AS districts
    FROM 'Bharat-Health-Service-Atlas/data/district_year.parquet'
    WHERE fy IN (2012, 2016, 2019, 2021) GROUP BY fy ORDER BY fy""").df().to_string(index=False))

  fy  inst_delivery  c_section  districts
2012          0.825      0.080        399
2016          0.939      0.113        618
2019          0.982      0.140        595
2021          0.987      0.165        703

With the datasets library: load_dataset("judicialmind/Bharat-Health-Service-Atlas", "hsdi_index", split="train") (configs district_year, hsdi_index, rhs_infrastructure, indicators).

Files

file	content
`data/district_year.parquet`	yearly sums of 15 HMIS items, with months reported for each (`<item>_months`), 8 ratio indicators, `reporting_completeness`, facility nil-reporting share (2012-20), `sources`
`data/hsdi_index.parquet`	HSDI components, their within-year percentile ranks (`pr_*`), `hsdi` (0-100), `hsdi_rank`, `hsdi_quintile` (5 = best)
`data/rhs_infrastructure.parquet`	sub-centres, PHCs, CHCs, sub-district and district hospitals per district and year; rural population; required facilities, shortfall, shortfall share, rural population per facility
`data/indicators.parquet`	indicator definitions and HMIS item codes in the 2008 and 2017 formats
`sources.parquet`	one row per source resource: provider, title, URL, attribution

Schema

fy is the financial-year start (2021 = FY2021-22, April 2021 to March 2022). Districts carry LGD codes (state_lgd, district_lgd) and names (state_name, district_name).

data/district_year.parquet:

column	meaning
`reporting_months`	months with any HMIS report in the year
`anc_registered`, `anc_first_trimester`, `institutional_deliveries`, `home_deliveries_sba`, `home_deliveries_non_sba`, `c_sections`, `live_births`, `newborns_weighed`, `low_birth_weight`, `fully_immunised`, `mcv1`, `ipd_headcount_midnight`, `opd_all`, `malaria_tests`, `malaria_positive`	yearly sum of the HMIS item (definitions and codes in `indicators.parquet`)
`<item>_months`	months in which the item was reported
`anc_first_trimester_share`, `institutional_delivery_share`, `c_section_rate`, `full_immunisation_per_birth`, `mcv1_per_birth`, `newborn_weighed_share`, `low_birth_weight_share`, `malaria_test_positivity`	ratio indicators (fractions, not percent)
`reporting_completeness`	`reporting_months` ÷ 12
`facility_nil_share`, `silent_unit_share`	mean monthly share of facilities filing a nil report, and of reporting units that filed no report (FY2012-20, from the HMIS reporting-status files)
`sources`	HMIS source(s) behind the row: `HMIS item-wise monthly reports`, `HMIS state reports (data.gov.in)`, `HMIS sub-district monthly file 2020-22`, or two of them joined with `+`

data/hsdi_index.parquet: the five components, pr_<component> (within-year percentile rank, 0-1), n_components, hsdi (0-100), hsdi_rank (1 = best in the year), hsdi_quintile (1-5, 5 = best).

data/rhs_infrastructure.parquet: sub_centres, phcs, chcs, sub_district_hospitals, district_hospitals, rural_population, rural_population_fy, and per facility type (sub_centres, phcs, chcs): <type>_required, <type>_shortfall, <type>_shortfall_share, rural_pop_per_<type>.

data/indicators.parquet: indicator, kind (count or ratio), label, hmis_2008_codes, hmis_2017_codes (for ratios, the formula). sources.parquet: uuid, source (data.gov.in, NDAP (NITI Aayog) or HMIS state reports (data.gov.in)), provider, title, url, attribution.

Results

Indicators (district medians)

FY	ANC in 1st trimester	institutional deliveries	C-section rate	fully immunised per birth	low birth weight
FY2012-13	0.552	0.825	0.080	1.049	0.116
FY2016-17	0.694	0.939	0.113	1.096	0.108
FY2019-20	0.792	0.982	0.140	1.116	0.112
FY2021-22	0.819	0.987	0.165	1.134	0.114

How the ratios are computed. Each ratio sums its numerator and denominator over only the months where both were reported.

Institutional delivery share = institutional ÷ (institutional + home deliveries).
Per birth ratios divide doses or children by reported live births. They exceed 1 where children are vaccinated in a district other than their district of birth, or where home births go unreported. Read them as service volume per birth, not as coverage.

Health Service Delivery Index (HSDI)

The HSDI is the equal-weight mean of within-year percentile ranks (0 = lowest district, 1 = highest) of 5 components, × 100. "Per birth" components are capped at 1 before ranking.

anc_first_trimester_share, institutional_delivery_share, full_immunisation_per_birth, newborn_weighed_share, reporting_completeness
A district-year is scored only if it reported at least 9 months and has at least 4 of the components.
The index is relative within a year: it ranks districts against each other, and does not compare a year with another.

Spearman correlation of each component with the index: anc_first_trimester_share 0.7, institutional_delivery_share 0.626, full_immunisation_per_birth 0.288, newborn_weighed_share 0.639, reporting_completeness 0.246. In FY2021-22, the top-scored districts are Cuddalore (80), Kaithal (80), Tirupathur (79), Ariyalur (79), Mayiladuthurai (79).

Rural Health Statistics

Facility counts come from NDAP dataset 7035 (Rural Health Statistics: Districtwise Health Care Infrastructure, Ministry of Health and Family Welfare). NDAP serves this dataset with coded column ids only. The names here (I7035_4→sub_centres, I7035_5→phcs, I7035_6→chcs, I7035_9→sub_district_hospitals, I7035_10→district_hospitals) were assigned by matching the national totals to the published RHS totals. Two further columns, present only in FY2018-19, are left out.

Rural population is the latest year of NDAP dataset 9307 (I9307_7, rural population, Jal Jeevan Mission).
Required facilities = rural population ÷ norm, using the IPHS plain-area norms: sub centres 1 per 5,000, phcs 1 per 30,000, chcs 1 per 120,000. Hilly and tribal areas have lower norms (3,000 / 20,000 / 80,000), so shortfalls there are understated.
Doctor and specialist vacancy tables exist only at state level in the portal, so they are not in this release.

Coverage

HMIS service indicators for 730 LGD districts, FY2009-10 to FY2021-22. District coverage grows over time: few districts before FY2012, about 550-600 in FY2013-19, about 700 from FY2020 with the 2020-22 sub-district file. sources names the HMIS source behind each district-year; values from the 2020-22 file run about 1.1× the item-wise reports where both exist, which shows as a level step at FY2020.
Person counts (deliveries, births, children) between 1 and 9 in a district-year are suppressed (null). Zeros are kept.
Districts are mapped to current LGD districts by name. A district split after 2009 maps to the successor that keeps the parent's name.
RHS facility counts for 734 districts, NDAP years labelled "Financial Year 2016" to "Financial Year 2021".
TB notification (NDAP 1139) is not included: its coded columns could not be matched to definitions.

Source data

Derived from 19,155 resources published on the Open Government Data Platform India (data.gov.in), snapshot May 2026, under the Government Open Data License - India. Per-resource attribution (provider, title, resource URL) is in the attribution column of sources.parquet. GODL-India does not cover personal information; the release holds district-level aggregate counts only.

Providers: Ministry of Health and Family Welfare (HMIS), Ministry of Health and Family Welfare (RHS, via NITI Aayog NDAP).

NDAP datasets 7035 and 9307: NITI Aayog National Data and Analytics Platform, https://ndap.niti.gov.in.

Original data: published by Ministry of Health and Family Welfare (Health Management Information System, HMIS; Rural Health Statistics, via NITI Aayog NDAP) on the Open Government Data Platform India (data.gov.in) and, where marked, NITI Aayog's National Data and Analytics Platform (NDAP), under the Government Open Data License - India (GODL-India). The original public data remains under GODL-India; the CC BY-NC-SA 4.0 licence of this repository covers JudicialMind's compilation, derived fields, annotations, benchmarks and models.

Citation

@misc{judicialmind2026healthatlas,
  title={Bharat-Health-Service-Atlas: District Health Service Delivery Indicators, HSDI and Rural Health Infrastructure for India},
  author={JudicialMind},
  year={2026},
  publisher={Hugging Face},
  howpublished={\url{https://huggingface.co/datasets/judicialmind/Bharat-Health-Service-Atlas}}
}

License

CC BY-NC-SA 4.0 (see LICENSE): research, academic, personal and other non-commercial use, with attribution to JudicialMind and to the original data publishers; adaptations must be shared under the same licence. Commercial use requires a separate licence.

Commercial use: contact surya@judicialmind.ai or hello@judicialmind.ai.

Downloads last month: 19

Total file size:

2.97 MB

Collection including judicialmind/Bharat-Health-Service-Atlas

India Open Government Data — Bharat Health Datasets

Collection

Research and non-commercial use (CC BY-NC-SA 4.0). Commercial: surya@judicialmind.ai / hello@judicialmind.ai • 2 items • Updated 1 day ago