This last blog entry is a summary of the work on our project. We examined the Twitter usage of German politicians throughout 2020 by analyzing their tweets in terms of Twitter-inherent measures as well as NLP- and SNA-related methods.

The topic proposal „Debate Structures around COVID-19 on Twitter“ aroused our interest. This topic was assigned to us and another group in the course. To delimit ourselves we decided after some research to investigate the Twitter usage of German politicians throughout 2020. For communication we mainly used Skype. Our meetings were usually 2 times a week, more often towards the end of the semester or for important dates.

To analyze the behavior of German politicians on Twitter, we used quantitative and content-based methods such as sentiment analysis and topic detection, as well as social network analysis methods. All these tools are used with the distinct features and characteristics of Twitter in mind.

The different types of analysis we performed give a holistic picture of the course of the pandemic in Germany throughout 2020. Our findings indicate that by tracking tweets of politicians, we can build an accurate timeline of important events, especially when they are related to COVID-19. While many politicians have a Twitter account, only some are highly active and influential. Their example shows that Twitter can be an effective communication tool, and more politicians should embrace this direct way of conveying information. Looking at the relations between the Twitter accounts of all politicians reveals, that political interaction is strongly shaped by party affiliations and the general political landscape. While the parties are strongly connected groups, their position to each other is far less divisive as in other countries such as the USA. We also find that parties are more divided regarding COVID-19 than ’everyday’ political issues. Content analysis of the politicians’ tweets shows, that the general sentiment displayed in them is rather negative, only slightly higher when regarding COVID-19. While non-COVID-tweets have a tendency to filled by anger, COVID-tweets more often express fear and tense anticipation of the future. It can also be noticed that tweets by politicians are more concerned with decision making (e.g. deciding that a lockdown is happening), rather than the consequences of these decisions (e.g. living conditions during a lockdown). The trend to get tired of the topic COVID-19 is generally felt throughout the population after dealing with the pandemic for over a year¹. This trend is also noticeable in the politicians tweets, indicating a developed routine in dealing with the virus. Connections of politicians’ Twitter accounts to those of public health experts are far weaker than expected. This presents an opportunity for both sides to create stronger publicly visible relations, in order to better communicate health measures to the public.

Through our findings we explain how politicians use Twitter as a means of communication during a crisis. The results may help politicians to more effectively use Twitter, and the general public to better understand government communication on social media.

Fortunately, the reviewers rated the paper as good, with only minor weaknesses, which could be corrected without problems. At this point, thanks to the reviewers for the constructive criticism.

If you are interested in the final paper, you can find it here: german_politicians_usage_of_twitter_during_the_covid_19_pandemic.pdf

¹https://projekte.uni-erfurt.de/cosmo2020/web/summary/34/