/covid19-analysis

Analysis of social media conversations surrounding COVID-19 in Singapore

Primary LanguageJupyter Notebook

Analysis of social media conversations surrounding COVID-19 in Singapore

The following files were used to analyse Facebook, Reddit and Twitter data collected from January 2020 to March 2020. As more data is collected, the code serves as a template to conduct further analysis. Analysis conducted include:

  • Topic modelling to identify key topics and user groups in Singapore’s social media scene
  • Sentiment analysis to model and evaluate netizen sentiments based on social media posts, against the growth of COVID-19 cases in Singapore
  • Social network analysis to model the Twitter retweet network in Singapore on posts related to COVID-19

A summary of the analysis results can be found here.

Brief summary of each file

Twitter Data Cleaning & Preprocessing.ipynb

  • Text pre-processing for Twitter data collected using the Twitter REST API

Retrieve Singapore Tweets For Analysis.ipynb

  • With data collected using Twitter REST API, filter tweets with location tag "SG" for analysis

Topic modelling - Twitter.ipynb

  • Conduct topic modelling (Latent Dirichlet Allocation) on data collected using Twitter REST API

Sentiment Analysis - Twitter .ipynb

  • Conduct sentiment analysis on Twitter data using VADER

Extract Twitter Network.ipynb

  • Extract Twitter retweet network in Singapore, for network analysis to be conducted (exports a gml file for visualisation on Gephi)

Topic Modelling - Reddit.ipynb

  • Conduct topic modelling (Latent Dirichlet Allocation) on data collected from Reddit API

Sentiment Analysis - Reddit.ipynb

  • Conduct sentiment analysis on Reddit data using VADER

Number of Facebook Posts Against Time.ipynb

  • Plot the number of Facebook posts against time for key ministers (of Singapore government) involved in COVID-19 response