---
license: odc-by
language:
- es
tags:
- economics
pretty_name: Minutas Banco de la República Colombia
size_categories:
- n<1K
task_categories:
- text-classification
formats:
- csv
- parquet
---


<span>
  <a href="https://huggingface.co/datasets/yabramuvdi/MinutasBanRepCol" style="display: inline-block; margin-right: 10px;">
    <img src="https://img.shields.io/badge/Dataset_Card_Español-0353A4" alt="Dataset Card Español">
  </a>
  <a href="https://huggingface.co/datasets/yabramuvdi/MinutasBanRepCol/blob/main/README_EN.md" style="display: inline-block;">
    <img src="https://img.shields.io/badge/Dataset_Card_English-0353A4" alt="Dataset Card English">
  </a>
</span>


## Description

This corpus was created to compile the monetary policy minutes of the Banco de la República (the Central Bank of Colombia) and make them easily accessible for research purposes. The minutes describe the factors considered by the Board members of the Banco de la República when making interest rate policy decisions, as well as the decisions themselves. Central bank minutes corpora have been widely used in economic literature is topics such as to build sentiment indicators, extract monetary policy indicators, and analyze economic variables from unstructured text data. Most existing corpora are available in English. However, to the best of our knowledge, this is the first corpus on monetary policy available in Spanish.

The corpus begins in June 2007, when the Board of Directors of the Banco de la República [started publishing the minutes](https://www.banrep.gov.co/es/node/6722).

The data can be loaded using Python as follows:

```python
from datasets import load_dataset

ds = load_dataset("yabramuvdi/MinutasBanRepCol", "default", split="train")

```

## Motivation

Economic literature has recently utilized central bank communications as a rich source of data for analysis. For example, [Hansen *et al. (2018)*](https://ora.ox.ac.uk/objects/uuid:a3818f48-ccb7-4ab8-be6b-d9fde5ac13c0/files/m885371aa70bb69ff66ebc28595e8a6ad), studied how transparency influenced monetary policy deliberations at the Federal Reserve.

However, there are currently no consolidated public data sources available for central banks whose official language is Spanish. With this corpus, we aim to pave the way for increasing the number of high-quality public datasets in Spanish.


## Structure

This database is in tabular format (185 x 2). Each record contains the following attributes:

1. **fecha**: The year, month, and day (YYYY-MM-DD) of the minute's publication. The day is not precise; thus, all dates use the first day of the month.
2. **text**: The text extracted from the original PDF minutes.

## Data Source

The original PDF files containing the minutes can be found on the [official website of the Banco de la República de Colombia](https://www.banrep.gov.co/es/minutas). Special thanks to Fredy Muñoz for his collaboration in consolidating the minutes dataset.

## Data Processing

The text from the minutes was extracted from the PDFs using [PyMuPDF](https://pymupdf.readthedocs.io/en/latest/). To preserve as much information as possible and leave relevant preprocessing steps to the user's discretion, no text preprocessing was applied.

## Usage Policy and Citation

The corpus is freely available for non-commercial purposes, provided attribution is given to the authors:


```bash
@misc{Muvdi2024BanRep
      title={Minutas Política Monetaria Banco de la República Colombia}, 
      author={Muvdi Yabra and Javier Gómez},
      year={2024},
      url = {https://huggingface.co/datasets/yabramuvdi/MinutasBanRepCol/}
}
```