SidneyBissoli/sipni-microdados

---
language:
- pt
license: cc-by-4.0
tags:
- health
- brazil
- public-health
- parquet
- datasus
- sipni
- vaccination
- immunization
pretty_name: "SI-PNI — Routine Vaccination Microdata (Brazil)"
size_categories:
- 100M<n<1B
task_categories:
- tabular-classification
source_datasets:
- original
---

# SI-PNI — Routine Vaccination Microdata (Brazil, 2020–present)

Individual-level vaccination records from Brazil's National Immunization
Program (SI-PNI), redistributed as partitioned Apache Parquet for efficient
analytical access. Each row represents one administered dose.

**Part of the [healthbr-data](https://huggingface.co/SidneyBissoli) project** — open redistribution of Brazilian public health data.

## Summary

| Item | Detail |
|------|--------|
| **Official source** | OpenDATASUS / Ministry of Health |
| **Temporal coverage** | January 2020 – present (monthly updates) |
| **Geographic coverage** | All 5,570 Brazilian municipalities |
| **Granularity** | Individual record (one row per administered dose) |
| **Volume** | 736M+ records |
| **Format** | Apache Parquet, partitioned by `ano/mes/uf` |
| **Data types** | All fields stored as `string` (preserves leading zeros) |
| **Update frequency** | Monthly |
| **License** | CC-BY 4.0 |

## Resumo em português

**SI-PNI — Microdados de Vacinação de Rotina (Brasil, 2020–presente)**

Microdados individuais de vacinação de rotina do Sistema de Informação do
Programa Nacional de Imunizações (SI-PNI), redistribuídos em formato Apache
Parquet particionado para acesso analítico eficiente. Cada linha representa
uma dose aplicada.

| Item | Detalhe |
|------|---------|
| **Fonte oficial** | OpenDATASUS / Ministério da Saúde |
| **Cobertura temporal** | Janeiro/2020 – presente (atualização mensal) |
| **Cobertura geográfica** | Todos os 5.570 municípios brasileiros |
| **Granularidade** | Registro individual (uma linha por dose aplicada) |
| **Volume** | 736M+ registros |
| **Formato** | Apache Parquet, particionado por `ano/mes/uf` |
| **Atualização** | Mensal |

> Para documentação completa em português, consulte o
> [repositório do projeto](https://github.com/SidneyBissoli/healthbr-data).

## 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)

```r
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/microdados/", format = "parquet")

# Example: vaccines administered in Acre, January 2024
ds |>
filter(ano == "2024", mes == "01", uf == "AC") |>
count(ds_vacina) |>
collect()
```

### Python (PyArrow)

```python
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/microdados/",
filesystem=s3,
format="parquet",
partitioning="hive"
)

table = dataset.to_table(
filter=(pds.field("ano") == "2024") & (pds.field("uf") == "AC")
)
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/microdados/
README.md
ano=2020/
mes=01/
uf=AC/
part-0.parquet
uf=AL/
part-0.parquet
...
mes=02/
...
ano=2021/
...
```

Each Parquet file contains records for a specific month and state.
Hive-style partitioning (`key=value`) enables automatic partition pruning
in Arrow and DuckDB — filtered queries read only the relevant files.

## Schema

The dataset contains 56 variables, all stored as `string` to preserve
leading zeros in IBGE municipality codes, CNES facility codes, ZIP codes,
and race/ethnicity codes. Key variables include:

| Variable | Description |
|----------|-------------|
| `dt_vacina` | Vaccination date (YYYY-MM-DD) |
| `co_vacina` | Immunobiological code |
| `ds_vacina` | Immunobiological description |
| `co_dose` | Dose code (1st, 2nd, 3rd, booster, etc.) |
| `sg_uf` | State abbreviation |
| `co_municipio_ibge` | IBGE municipality code (6 digits) |
| `co_cnes` | CNES health facility code |
| `dt_nascimento` | Patient date of birth |
| `co_sexo` | Sex (M/F) |
| `co_raca_cor` | Self-reported race/ethnicity |

> For the complete 56-variable data dictionary, see the Ministry of Health's
> `Dicionario_tb_ria_rotina.pdf`.

## Source and processing

**Original source:** Compressed JSON files from OpenDATASUS (Ministry of
Health S3 bucket).

**Why JSON instead of CSV?** The CSV exports from 2020–2024 contain
serialization artifacts (numeric fields with `.0` suffix, loss of leading
zeros). JSON preserves all values as strings with full integrity.

**Processing:** JSON → NDJSON (via `jq`) → Parquet (via `polars`) → upload
to R2 (via `rclone`). No transformations are applied — values are published
exactly as provided by the Ministry of Health.

## Known limitations

1. **Government data, not ours.** Errors in the original data are
intentionally preserved. No cleaning or correction is applied.
2. **Variable completeness.** Many fields have optional reporting and may
contain high proportions of empty values or "SEM INFORMACAO".
3. **All fields are strings.** Type casting (Date, integer) must be done by
the user at analysis time.
4. **Temporal coverage.** Individual-level microdata is available only from
January 2020. For the 1994–2019 historical series, see the aggregated
datasets: `sipni-agregados-doses` and `sipni-agregados-cobertura`.
5. **Lag.** The Ministry may take weeks to publish a given month's data.
The pipeline runs monthly and reflects what is available at the source.
6. **Does not include COVID-19.** COVID vaccination data is in a separate
dataset: `sipni-covid`.

## Citation

```bibtex
@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-microdados},
note = {Original source: Ministry of Health / OpenDATASUS}
}
```

## Contact

- **GitHub:** [https://github.com/SidneyBissoli](https://github.com/SidneyBissoli)
- **Hugging Face:** [https://huggingface.co/SidneyBissoli](https://huggingface.co/SidneyBissoli)
- **E-mail:** [e-mail]

---

*Last updated: 2026-02-28*

语言：
- 葡萄牙语
许可协议：CC-BY 4.0
标签：
- 健康
- 巴西
- 公共卫生
- Apache Parquet
- OpenDATASUS
- SI-PNI
- 疫苗接种
- 免疫接种
美观名称："SI-PNI——巴西常规疫苗接种微数据"
数据规模分类：
- 1亿至10亿条记录
任务类别：
- 表格分类
源数据集：
- 原始数据集
---

# SI-PNI——巴西常规疫苗接种微数据（2020年至今）

个体级疫苗接种记录源自巴西国家免疫计划（National Immunization Program, SI-PNI），经重新分发为分区存储的Apache Parquet格式，以支持高效的分析访问。每一行对应一剂已接种的疫苗。

**本数据集隶属于[healthbr-data](https://huggingface.co/SidneyBissoli)项目**——该项目致力于开放重分发巴西公共卫生数据。

## 数据概览

| 项目 | 详情 |
|------|--------|
| **官方来源** | OpenDATASUS / 巴西卫生部 |
| **时间覆盖范围** | 2020年1月至今（每月更新） |
| **地理覆盖范围** | 巴西全部5570个直辖市 |
| **数据粒度** | 个体记录（每一行对应一剂已接种疫苗） |
| **数据体量** | 7.36亿+条记录 |
| **数据格式** | Apache Parquet，按`ano/mes/uf`分区 |
| **数据类型** | 所有字段均存储为字符串（保留前导零） |
| **更新频率** | 每月 |
| **许可协议** | CC-BY 4.0 |

## 葡萄牙语摘要（已翻译）

**SI-PNI——巴西常规疫苗接种微数据（2020年至今）**

本数据集源自巴西国家免疫计划信息系统（SI-PNI）的个体级常规疫苗接种微数据，经重新分发为分区存储的Apache Parquet格式，以支持高效的分析访问。每一行对应一剂已接种的疫苗。

| 项目 | 详情 |
|------|--------|
| **官方来源** | OpenDATASUS / 巴西卫生部 |
| **时间覆盖范围** | 2020年1月至今（每月更新） |
| **地理覆盖范围** | 巴西全部5570个直辖市 |
| **数据粒度** | 个体记录（每一行对应一剂已接种疫苗） |
| **数据体量** | 7.36亿+条记录 |
| **数据格式** | Apache Parquet，按`ano/mes/uf`分区 |
| **更新频率** | 每月 |

> 如需查看完整的葡萄牙语文档，请参阅项目[代码仓库](https://github.com/SidneyBissoli/healthbr-data)。

## 数据访问

数据集托管于Cloudflare R2存储服务，通过兼容S3的API进行访问。以下提供的凭据为**只读权限**，仅供公共使用。

### R语言（Arrow库）

r
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/microdados/", format = "parquet")

# 示例：查询2024年1月阿克里州的疫苗接种量
ds |>
filter(ano == "2024", mes == "01", uf == "AC") |>
count(ds_vacina) |>
collect()

### Python（PyArrow库）

python
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/microdados/",
filesystem=s3,
format="parquet",
partitioning="hive"
)

table = dataset.to_table(
filter=(pds.field("ano") == "2024") & (pds.field("uf") == "AC")
)
print(table.to_pandas().head())

> **注意：** 这些凭据为只读权限，可安全用于脚本中。该存储桶不支持匿名S3访问，必须使用凭据方可访问。

## 文件结构

s3://healthbr-data/sipni/microdados/
README.md
ano=2020/
mes=01/
uf=AC/
part-0.parquet
uf=AL/
part-0.parquet
...
mes=02/
...
ano=2021/
...

每个Parquet文件存储特定月份和州的接种记录。采用Hive风格的分区格式（`键=值`），可在Arrow和DuckDB中实现自动分区裁剪——过滤后的查询仅会读取相关文件。

## 数据模式

本数据集共包含56个变量，所有字段均以字符串类型存储，以保留IBGE（巴西地理与统计研究所）直辖市代码、CNES（巴西国家医疗机构登记系统）医疗机构代码、邮政编码以及种族/族裔代码中的前导零。核心变量如下：

| 变量 | 说明 |
|----------|-------------|
| `dt_vacina` | 疫苗接种日期（格式为YYYY-MM-DD） |
| `co_vacina` | 免疫生物制剂代码 |
| `ds_vacina` | 免疫生物制剂说明 |
| `co_dose` | 剂次代码（首剂、第二剂、第三剂、加强针等） |
| `sg_uf` | 州缩写 |
| `co_municipio_ibge` | IBGE直辖市代码（6位数字） |
| `co_cnes` | CNES医疗机构代码 |
| `dt_nascimento` | 接种者出生日期 |
| `co_sexo` | 性别（M/F） |
| `co_raca_cor` | 自我报告的种族/族裔 |

> 如需查看完整的56变量数据字典，请参阅巴西卫生部发布的`Dicionario_tb_ria_rotina.pdf`。

## 源数据与处理流程

**原始数据源：** 来自OpenDATASUS（巴西卫生部S3存储桶）的压缩JSON文件。

**为何选择JSON而非CSV？** 2020年至2024年的CSV导出文件存在序列化异常（数值字段带有`.0`后缀、丢失前导零等问题）。JSON可将所有值以字符串形式存储，完整保留数据完整性。

**处理流程：** 经`jq`工具将原始JSON文件转换为NDJSON（换行分隔JSON格式），再通过`polars`库转换为Parquet格式，最后通过`rclone`上传至R2存储服务。未对数据进行任何转换——所有值均严格按照巴西卫生部提供的原始内容发布。

## 已知局限性

1. **数据源自政府官方，非本项目所有。** 原始数据中的错误均被保留，未进行任何清洗或修正操作。
2. **变量完整性。** 多数字段为可选上报项，可能包含大量空值或“SEM INFORMACAO”（无信息）字段。
3. **所有字段均为字符串类型。** 需由使用者在分析阶段自行进行类型转换（如日期、整数类型）。
4. **时间覆盖范围。** 个体级微数据仅可追溯至2020年1月。如需1994年至2019年的历史序列数据，请参阅聚合数据集`sipni-agregados-doses`与`sipni-agregados-cobertura`。
5. **数据延迟。** 巴西卫生部可能需要数周时间才能发布当月数据。本项目的处理流程按月运行，数据内容严格反映源端当前可获取的内容。
6. **不含新冠疫苗接种数据。** 新冠疫苗接种数据已单独发布至数据集`sipni-covid`。

## 引用格式

bibtex
@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-microdados},
note = {Original source: Ministry of Health / OpenDATASUS}
}

## 联系方式

- **GitHub:** [https://github.com/SidneyBissoli](https://github.com/SidneyBissoli)
- **Hugging Face:** [https://huggingface.co/SidneyBissoli](https://huggingface.co/SidneyBissoli)
- **电子邮箱:** [e-mail]

---

*最后更新时间：2026年2月28日*