Mehdi009's picture
Create README.md
7e25ab5 verified
metadata
dataset_name: Antisemitism Harassment Detection Dataset
tags:
  - antisemitism
  - harassment-detection
  - audio-classification
  - text-classification
  - social-media
  - csv
task_categories:
  - text-classification
  - audio-classification
language:
  - en

Dataset Card for Antisemitism Harassment Detection Dataset

Dataset Summary

This dataset contains labeled examples of tweets related to antisemitism harassment. It includes the tweet content, user information, and labels indicating whether the tweet contains biased or harassment content. The dataset can be used for training and evaluating models in the domain of antisemitism harassment detection, text classification, and sentiment analysis.

Supported Tasks

The dataset can be used for the following tasks:

  • Text Classification: Detecting antisemitism or biased content in social media posts.
  • Audio Classification: Detecting antisemitism harassment in spoken audio (if audio data is generated).

Languages

The language of the dataset is English (en).

Dataset Structure

Columns

The dataset contains the following columns:

  • TweetID: A unique identifier for each tweet.
  • Username: The username of the person who posted the tweet.
  • Text: The content of the tweet.
  • CreateDate: The date and time the tweet was posted.
  • Biased: A binary label indicating whether the tweet is biased (1) or not (0).
  • Keyword: The keyword associated with the tweet content.

Example Row

TweetID Username Text CreateDate Biased Keyword
1.23e18 Celtic_Films AIPAC should be registered as a foreign agent ... 2020-02-15 17:57:21+00 1 Israel

Usage

The dataset can be used for training models to detect antisemitism and harassment in social media posts. It can also be extended to audio classification by converting text data into synthesized speech and training audio models like Wav2Vec2 or Hubert.

Load Dataset

To load the dataset in Python, use the following code:

from datasets import load_dataset

# Load the dataset from a local CSV file
dataset = load_dataset("csv", data_files={"train": "path/to/GoldStanderDataSet.csv"})

# Inspect the first few rows
print(dataset["train"].head())