Twitter Bot for Disaster Management & Relief
- 0 Collaborators
A NLP based bot on Twitter which utilizes methodologies like Text Summarization, Text Classification & Automated Twitter Streamings for Disaster Management Related tasks. ...learn more
Project status: Under Development
Networking, Artificial Intelligence
Groups
Student Developers for AI
Overview / Usage
This Twitter Bot will collect tweets related to earthquakes and similar crisis around the world. It'll extract information and classify it into 4 sub-categories: Damage to Human Life, Damage to Infrastructure, Volunteers & Relief and Unrelated information. It'll further summarize this information and share it on Twitter as a public alert/
Methodology / Approach
The tweets are collected using Twitter's streaming API. The tweets are further pre-processed, cleaned and classified into 4 categories using feature extraction techniques like TF-IDF vectors and Word Embeddings. The classification model is trained on various publicly available datasets. The classified tweets are summarized using a Encoder-Decoder Network and finally posted on Twitter for public information.
Technologies Used
Twitter Streaming API, Keras, NLTK, SpaCy.