Twitter has recently become one of the most important means of communication when an emergency takes place. The use of smartphones allows people to report an emergency that is occurring in real time. Therefore, there is a marked interest in monitoring Twitter and knowing how to automatically predict whether a real emergency or a common situation is being faced.
This guide reveals an approach based on Machine Learning (ML) that combines the expertise of Natural Language Processing (NLP) techniques and classification methods. The aim is to provide an automatic tool for addressing a binary classification problem in disaster prediction.
Dataset: NLP prediction of Disaster Tweets
Code: Colab notebook