This repository hosts an ongoing project focused on applying Natural Language Processing (NLP) techniques to assess flood severity in the UK using articles scraped from the FloodList website. The main objectives of this project are:
Data Collection:
A dataset of 80 flood-related articles from the UK has been collected through web scraping.
Dataset Preparation:
Manual labeling of 20 articles to create an initial training set.
Training a model on 40 articles to predict flood severity levels.
Testing the model on the remaining 20 articles.
NLP Application:
Develop and implement NLP techniques to analyze article content.
Build a classification model to identify the severity of floods described in the articles.
This repository currently contains:
Scraping scripts used to collect the articles.
Preliminary datasets.
Early-stage NLP code.
Next Steps:
Improve dataset labeling and preprocessing.
Design and train an initial model for severity classification.
Evaluate model performance and refine methods.
Note: This project is a work in progress. Contributions and feedback are welcome! 🚀