#GrenfellTowerFire
Apoorva, Olga,
Justin, Steffie, Yimei
Mentor: Dr Miriam Fernandez
Slides Available:
https://justinchuntingho.github.io/CSS-slides/
2017 Grenfell Tower Fire
WHERE:
Grenfell Tower, North Kensington, London, UK
WHEN:
14 June 2017, 00:54 Local Time
WHAT:
60 hours of fire until extinguished
CASUALTIES:
72 died, over 70 injured, 2000 lost homes
Research Questions
RQ1:
How did the public sentiment about the Grenfell Tower Fire on Twitter evolve over time?
RQ2:
How did the conversation about the Grenfell Tower Fire on Twitter evolve over time?
Methods
Data: Sub-sample (10%) of 145k Tweets
Sentiment Analysis
Topic Modeling
Sentiment Analysis
TextBlob vs. Vader
TextBlob vs. Vader
Topic Modeling
Topic Modeling
+
Sentiment Analysis
Conclusion
1. People sleep at night
2. Big data is too big
3. If you are happy when tragedy happens, you are probably not human
4. The Queen always wins in the end
Real Conclusion
1. We see overall negative sentiment for tweets after crisis like this.
2. Sentiment is context-dependent. Different pattern of trends can only be seen once we look at specific topic.
Implication and Impact:
1. Twitter can be a useful platform for public communication during and after a crisis.
2. Announcement of support from government and organisations would generate positive sentiment on social media after crisis.
Thank You!
