Tweets_Dataset_for_Sarcasm_detection_in_Hindi

URL: https://github.com/pragyakatyayan/Tweets_Dataset_for_Sarcasm_detection_in_Hindi

Description

This repository contains a raw dataset of over 16,000 tweets, including both sarcastic and non-sarcastic examples, aimed at researchers working on sarcasm detection in Hindi. The dataset includes tweets in native Hindi and is particularly focused on specific hashtags. The data was collected using a scraping script.

Methods

The dataset can be obtained via two methods:

1. Scraping with Python Script: Researchers can use the scrap_tweets_in_Hindi-v1.py file to rescrape tweets directly from Twitter.
2. Download the Jupyter Notebook: The dataset can also be downloaded and used by running the Jupyter Notebook provided in the repository.

The data collection was done using pandas, tweepy, and textblob.

Results

The dataset contains:

• 6051 Sarcastic Tweets
• 10128 Non-Sarcastic Tweets

The dataset spans the period from 01-01-2012 to 23-06-2020.

Models

No specific models are provided in this repository. However, the dataset is meant for use in sarcasm detection models for Hindi-language tweets.

Dataset

The dataset includes:

• Sarcastic tweets (6051)
• Non-sarcastic tweets (10128)

Tweets were scraped using the code from Mr. Griffin Leow’s GitHub repository, with modifications to extract tweets in Hindi.