metadata

annotations_creators:
  - crowdsourced
language_creators:
  - crowdsourced
language:
  - pt
license:
  - cc-by-sa-4.0
multilinguality:
  - monolingual
size_categories:
  - 10K<n<100K
source_datasets:
  - original
task_categories:
  - text-classification
task_ids: []
paperswithcode_id: told-br
pretty_name: ToLD-Br
language_bcp47:
  - pt-BR
tags:
  - hate-speech-detection
dataset_info:
  - config_name: multilabel
    features:
      - name: text
        dtype: string
      - name: homophobia
        dtype:
          class_label:
            names:
              '0': zero_votes
              '1': one_vote
              '2': two_votes
              '3': three_votes
      - name: obscene
        dtype:
          class_label:
            names:
              '0': zero_votes
              '1': one_vote
              '2': two_votes
              '3': three_votes
      - name: insult
        dtype:
          class_label:
            names:
              '0': zero_votes
              '1': one_vote
              '2': two_votes
              '3': three_votes
      - name: racism
        dtype:
          class_label:
            names:
              '0': zero_votes
              '1': one_vote
              '2': two_votes
              '3': three_votes
      - name: misogyny
        dtype:
          class_label:
            names:
              '0': zero_votes
              '1': one_vote
              '2': two_votes
              '3': three_votes
      - name: xenophobia
        dtype:
          class_label:
            names:
              '0': zero_votes
              '1': one_vote
              '2': two_votes
              '3': three_votes
    splits:
      - name: train
        num_bytes: 2978006
        num_examples: 21000
    download_size: 2430416
    dataset_size: 2978006
  - config_name: binary
    features:
      - name: text
        dtype: string
      - name: label
        dtype:
          class_label:
            names:
              '0': not-toxic
              '1': toxic
    splits:
      - name: train
        num_bytes: 1709560
        num_examples: 16800
      - name: test
        num_bytes: 216297
        num_examples: 2100
      - name: validation
        num_bytes: 212153
        num_examples: 2100
    download_size: 853322
    dataset_size: 2138010

Portuguese Hate Speech Dataset (TuPy)

The Portuguese hate speech dataset (TuPy) is an annotated corpus designed to facilitate the development of advanced hate speech detection models using machine learning (ML) and natural language processing (NLP) techniques. TuPy is formed by 10000 thousand unpublished annotated tweets collected in 2023.

This repository is organized as follows:

root.
    ├── annotations   : classification given by annotators
    ├── raw corpus    : dataset before being split between annotators
    ├── tupy datasets : combined result of annotations
    └── README.md

Voting process

To generate the binary matrices, we employed a straightforward voting process. Three distinct evaluations were assigned to each document. In cases where a document received two or more identical classifications, the adopted value is set to 1; otherwise, it is marked as 0.

Acknowledge

The TuPy project is the result of the development of Felipe Oliveira's thesis and the work of several collaborators. This project is financed by the Federal University of Rio de Janeiro (UFRJ) and the Alberto Luiz Coimbra Institute for Postgraduate Studies and Research in Engineering (COPPE).