# A geocoded dataset of primary health care clinics in Brazil

## TL;DR

This paper creates a detailed dataset of primary health care clinic locations in Brazil using publicly available data sources.

## Contribution

The novel contribution is a geocoded dataset of primary care clinics in Brazil, enabling spatial analysis of health care access.

## Key findings

- The dataset includes monthly observations for 293,698 clinics from January 2018 to December 2023.
- Geographic coordinates are assigned using postal codes from the IBGE-CNEFE registry.
- The method's precision is evaluated by estimating postal code areas using IBGE shapefiles.

## Abstract

We develop a geocoded dataset of primary health care clinics in Brazil. We merge data from three publicly available sources. The first is the National Registry of Healthcare Facilities (CNES-ST), which collects the location (state, municipality, and 8-digit postal code) of all health care facilities, public or private, operating in Brazil. The second is the National Registry of Addresses for Statistical Purposes (IBGE-CNEFE), which contains the geographic coordinates of all addresses in Brazil (including 8-digit postal codes) and serves as the basis for the Brazilian census. Our approach aggregates individual (address-level) coordinates to the 8-digit postal code, and assigns coordinates to primary care clinics based on each clinics’ postal code. Using data from a third source, the IBGE shapefiles, we estimate the area of postal codes to evaluate the precision of our geo-referencing method. The unique facility identification number (cnes number) can be used to merge our georeferenced data with other publicly available databases of the Brazilian Unified Health System. The final dataset is an unbalanced panel with monthly observations about 293,698 primary care clinics’ locations (i.e. coordinates), from January 2018 to December 2023, totalling 15,455,219 observations.