Future Research and Limitations

Although this project was largely successful, it could certainly be improved. Below are some ideas for areas of future research and possible improvements. 

• Hand coding a dataset to establish a clear ground truth to further establish validity of methods used. 
•  Expanding the dataset to include additional news sources would give a more robust picture of the coverage surrounding different newsworthy events. 
•  Disparate news source could be used to assess how narratives of the same event differ based on the country or news entity providing the coverage.   
• Introducing secondary sources such as social media content and less formalized text could yield interesting results and help determine how information regarding events spreads.    
•  The run times for fully processing the data were lengthy.  
•  It would be beneficial to find methods to speed up the processing time, possibly by using faster algorithmns for the clustering and extraction portions or by engineering the code to accommocate increased threading and GPU processing.
•  Building out the text standardization for the elements returned by the “Giveme5W1H” algorithm would also help faciitate the network analysis of narrative chains and make these methods more robust.