SidneyBissoli/sipni-agregados-cobertura

SI-PNI — Aggregated Vaccination Coverage (Brazil, 1994–2019)

Historical aggregated vaccination coverage data from Brazil's National Immunization Program (SI-PNI), covering 26 years of municipality-level coverage indicators pre-calculated by the Ministry of Health. Converted from legacy .dbf files to Apache Parquet.

Part of the healthbr-data project — open redistribution of Brazilian public health data.

Summary

Item	Detail
Official source	DATASUS FTP / Ministry of Health
Temporal coverage	1994–2019
Geographic coverage	All Brazilian municipalities (by state)
Granularity	Aggregated: one row per municipality × composite vaccine indicator
Volume	2.8M+ records (686 .dbf files processed)
Format	Apache Parquet, partitioned by ano/uf
Data types	All fields stored as string (preserves original format)
Update frequency	Static (historical series, no longer updated at source)
License	CC-BY 4.0

Resumo em português

SI-PNI — Cobertura Vacinal Agregada (Brasil, 1994–2019)

Dados históricos agregados de cobertura vacinal do Programa Nacional de Imunizações (PNI), cobrindo 26 anos de indicadores de cobertura em nível municipal, pré-calculados pelo Ministério da Saúde. Convertidos de arquivos .dbf legados para Apache Parquet.

Item	Detalhe
Fonte oficial	FTP DATASUS / Ministério da Saúde
Cobertura temporal	1994–2019
Cobertura geográfica	Todos os municípios brasileiros (por UF)
Granularidade	Agregado: uma linha por município × indicador composto de vacina
Volume	2,8M+ registros (686 arquivos .dbf processados)
Formato	Apache Parquet, particionado por ano/uf
Atualização	Estática (série histórica, não atualizada na fonte)

Para documentação completa em português, consulte o repositório do projeto.

Data access

Data is hosted on Cloudflare R2 and accessed via S3-compatible API. The credentials below are read-only and intended for public use.

R (Arrow)

library(arrow)
library(dplyr)

Sys.setenv(
  AWS_ENDPOINT_URL      = "https://5c499208eebced4e34bd98ffa204f2fb.r2.cloudflarestorage.com",
  AWS_ACCESS_KEY_ID     = "28c72d4b3e1140fa468e367ae472b522",
  AWS_SECRET_ACCESS_KEY = "2937b2106736e2ba64e24e92f2be4e6c312bba3355586e41ce634b14c1482951",
  AWS_DEFAULT_REGION    = "auto"
)

ds <- open_dataset("s3://healthbr-data/sipni/agregados/cobertura/", format = "parquet")

# Example: coverage indicators in São Paulo, 2015
ds |>
  filter(ano == "2015", uf == "SP") |>
  head(20) |>
  collect()

Python (PyArrow)

import pyarrow.dataset as pds
import pyarrow.fs as fs

s3 = fs.S3FileSystem(
    endpoint_override="https://5c499208eebced4e34bd98ffa204f2fb.r2.cloudflarestorage.com",
    access_key="28c72d4b3e1140fa468e367ae472b522",
    secret_key="2937b2106736e2ba64e24e92f2be4e6c312bba3355586e41ce634b14c1482951",
    region="auto"
)

dataset = pds.dataset(
    "healthbr-data/sipni/agregados/cobertura/",
    filesystem=s3,
    format="parquet",
    partitioning="hive"
)

table = dataset.to_table(
    filter=(pds.field("ano") == "2015") & (pds.field("uf") == "SP")
)
print(table.to_pandas().head())

Note: These credentials are read-only and safe to use in scripts. The bucket does not allow anonymous S3 access — credentials are required.

File structure

s3://healthbr-data/sipni/agregados/cobertura/
  README.md
  ano=1994/
    uf=AC/
      part-0.parquet
    uf=AL/
      part-0.parquet
    ...
  ano=1995/
    ...

Structural eras

The .dbf files underwent one major structural transition:

Era	Period	Columns	Key difference
1	1994–2012	9	Includes DOSE, FX_ETARIA; COBERT as numeric (decimal point)
2	2013–2019	7	DOSE and FX_ETARIA removed; COB as character (decimal comma)

The coverage field name and format changed: COBERT (numeric, periods) in era 1 vs COB (character, commas) in era 2. Both are preserved as-is.

Schema

Key variables (varies by era):

Variable	Description	Available
MUNICIP	Municipality code	All eras
IMESSION	Composite vaccine indicator code (per IMUNOCOB.DBF, 26 indicators)	All eras
COBERT / COB	Coverage percentage (pre-calculated by Ministry)	Era 1 / Era 2
QT_DOSE	Number of administered doses	All eras
POP	Target population (denominator)	All eras
DOSE	Dose type	Era 1 only
FX_ETARIA	Age group	Era 1 only

Important: The vaccine codes in coverage files use the IMUNOCOB.DBF dictionary (26 composite indicators), which is different from the IMUNO.CNV dictionary used in the doses files (85 individual vaccines). Coverage indicators often combine multiple individual vaccines into a single metric (e.g., "Polio" coverage combines OPV and IPV doses).

Source and processing

Original source: 702 .dbf files (dBase III) from the DATASUS FTP server. Of these, 686 were successfully processed (remaining were unavailable or empty). Bootstrap time: 44 minutes for 2,762,327 records.

Processing: .dbf → R (foreign::read.dbf) → Parquet (arrow::write_dataset) → upload to R2 (rclone). No transformations are applied. Consolidated files (UF, BR, IG prefixes) were excluded.

Reproducibility & provenance

This is a static, closed series (source no longer updated by the Ministry), converted once by pipeline version 1.0.0. Provenance is recorded per partition in manifest.json on R2 (source URL, MD5 and size, processing timestamp, output file paths and sizes, record count) and per source file in data/controle_versao_sipni_agregados_cobertura.csv in the GitHub repository (MD5, size, record and column counts). The Parquet files themselves do not carry the embedded healthbr metadata record used by the dynamic datasets; use manifest.json to check size and record count of a copy and to locate the original .dbf on the DATASUS FTP. The pipeline is a deterministic function of the source (all fields read as text, no value transformation), so any partition can be regenerated from the original file and the pipeline script at the commit recorded in the repository history. Data are not versioned: the R2 copy is the latest publication; revisions by the Ministry replace files, and the history of what was published when lives in the version-control CSV's git log. Full policy and audit recipe: docs/policy-reproducibility-pt.md.

Known limitations

1. Government data, not ours. Values are preserved exactly as in the original .dbf files, including the pre-calculated coverage percentages.
2. Two structural eras. Dose and age group columns disappear in 2013. Coverage field name and decimal format change between eras.
3. Composite indicators. The IMUNOCOB dictionary combines multiple vaccines into single coverage metrics. The mapping rules are complex and changed over time.
4. All fields are strings. The coverage percentage field must be parsed by the user (note the decimal point vs comma difference between eras).
5. Static dataset. No longer updated at source after 2019.
6. Coverage ≠ doses. This dataset contains pre-calculated coverage rates. For raw dose counts, see sipni-agregados-doses.

Citation

@misc{healthbrdata,
  author = {Sidney da Silva Bissoli},
  title  = {healthbr-data: Redistribution of Brazilian Public Health Data},
  year   = {2026},
  url    = {https://huggingface.co/datasets/SidneyBissoli/sipni-agregados-cobertura},
  note   = {Original source: Ministry of Health / DATASUS}
}

Contact

• GitHub: https://github.com/SidneyBissoli
• Hugging Face: https://huggingface.co/SidneyBissoli
• E-mail: sbissoli76@gmail.com

---

Last updated: 2026-02-28