A PyTorch implementation of "Graph Convolutional Networks for Text Classification." (AAAI 2019)

Node Embeddings in Dynamic Graphs

Reproduction material for "#Secim2023: First Public Dataset for Studying Turkish General Election"

Twitter Dynamic Dataset Api. Create any dataset YOU want.

[NLP] Unsupervised User Stance Detection on Twitter.

Temporal Walk Based Centrality Metric for Graph Streams

A collection of multiple social media dataset samples. Each sample contains over 1,000 records. These datasets are ideal for brand awareness, consumer sentiment analysis, and for tracking social media presence

Different bangla datasets for sentiment analysis on bangla text

Classifying users on social media, using text embeddings from OpenAI and others

Retrieve tweets from Twitter API with tweet ID

User Profiling and Sentiment analysis of Twitter social network during the impeachment of Brazilian President

The files in this repo were used in the analysis for the paper Branching process descriptions of information cascades on Twitter.

Utility python package for RG17 and UO17 Twitter tennis data sets.

Sentiment Analysis on Twitter Dataset

A Farsi (Persian) Semantic Similarity Measurement Dataset (FarSSiM)

Suicide Ideation Detection in Indonesian Language Twitter Posts Using IndoBERT and LSTM/CNN

End-to-end unsupervised text analytics pipeline processing 390K+ #MeToo tweets with Python, including regex-based text normalization, vectorization, topic modeling (LDA/NMF), sentiment scoring (TextBlob), and temporal pattern mining. Provides interpretable topic clusters and visualization of discourse trajectories during the movement.

This research uses Twitter data to develop a content analysis model, evaluating multiple machine learning models and interpreting the best one with LIME. A bi-directional LSTM with word2vec embeddings will also be applied, with performance compared against traditional ML models.

Exploring the Australian Energy generation data set and the twitter data set using Python

複数のソーシャルメディアデータセットサンプルのコレクションです。各サンプルには1,000件を超えるレコードが含まれています。これらのデータセットは、ブランド認知、消費者センチメント分析、およびソーシャルメディアでのプレゼンス追跡に最適です。