Honours_Final_Code

Thesis Title: Location extraction from Twitter text: A comparison of NLP and Machine Learning methods on Australian Natural Disaster Datasets

This repository is split into:

1. Data collection code.
   • Twitter_streamer.py
   • Twitter
2. Data preprocessing + filtering code
   • An example of preprocessing code for the Melbourne dataset.
3. Using Geograpy3
   • geograpy3_Melbourne.py
   • geograpy3_Seroja.py
4. Manual annotation and calculation of F1 measure.
   • Confusion_matrix_Melb.py
   • Confusion_matrix_seroja.py
   • Melb_geograpy3_random100_F1_measure
   • Seroja_geograpy3_F1_measure_Method1_2

Due to Twitter API Guidelines, Tweets can not be shared. For access to datasets, please contact k.rome@student.unsw.edu.au