In the last article, we talked in depth about how we are building the corpus for our analysis. As the corpus grows, we are also thinking about fitting methods to retrieve the results we want to. As already stated, we are focussing on analyzing the debate and conversation structure around the topic COVID-19 and more specific conversations that are startet by the 20 accounts whose tweets we are scraping.

For that reason, we need several methods to combine distant, as well as close reading and to get an overview of the corpus, as well as insight into specific conversations. In general, we want to do research on three different aspects of the dataset.

1) Look into the general topics, linguistic structure and statistics of the corpus (e.g. do Topic Modelling, Named Entity Recognition and such)

2) Do a network analysis to get a (visual) overview of how conversations are structured in our dataset and what drives them forward

3) Combine the results of the previous sections, as well as qualitative assessment to analyze conversations in depth (e.g. compare different kinds of first reactions to a tweet, see how the sentiment changes over time)

With these three section, we are confident to explore the dataset and generate interesting results. We are applying those methods already and at the moment deciding which concrete scripts we still need to run and how we can combine those results into a good paper. Stay tuned!