Nane Kratzke



Twitter and the German Federal Elections

Published: 01 Feb 2017 (latest update: 12 Feb 2020)
Duration: 1st Feb. 2017 - 28th Feb. 2018
Funding: internal project (Lübeck University of Applied Sciences)

Research context

Data-driven political campaigns can be successful. “The Obama 2012 campaign used data analytics and the experimental method to assemble a winning coalition vote by vote. In doing so, it overturned the long dominance of TV advertising in U.S. politics and created something new in the world: a national campaign run like a local ward election, where the interests of individual voters were known and addressed.” Technology Review

But four years later, Hillary Clinton’s data-driven campaign organized by the same party failed under the eyes of the world. The question is why data-driven campaigns worked for Barack Obama but not for Hillary Clinton? There is an interesting article proclaiming that ‘data-driven’ campaigns are killing the US Democratic Party, because the wrong lessons from Obama’s success have been learned. Dave Gold states that “Democrats have allowed microtargeting to become microthinking. Each cycle, we speak to fewer and fewer people and have less and less to say” although adressing the right audience. So, whether this is true or not can not be answered by this project. However, it should be obvious that data collected accompanying such election campaigns might contain worthful insights.

Especially Twitter analysis of US election campaigns are done for a while. However, there exist only some open accessible Twitter datasets with a clear focus on political election campaigns in countries of the European Union. That is why this project records Twitter interactions for one further European country (Germany).

Research objectives

One major motivation to record this dataset is to collect Twitter data in the “hot” (pre-)phase of political election campaigns in Germany. Social media channels are used more or more intensively by all political parties to distribute political messages and it is more than likely that the professionality – or the data-driveness – will increase in the future. So, this dataset might be one of the latest recordable datasets without being affected by too much microtargeting effects in political campaigning in Germany. Therefore, it might be used for future studies analyzing the campaigns for the 19th German Bundestag but also as a reference dataset to measure ‘data-driven’ effects of political campaigns and microtargeting on Twitter in future campaigns.

  • Record a representative dataset of Twitter interactions during to pre- and hot-phase for the 19th German Bundestag elections.
  • Provide this dataset via an Open Data platform like Zenodo.
  • Develop or contribute to pragmatic software tools that enable to record Twitter datasets over long period of times.
  • Perform a network analysis of the recorded dataset regarding political parties in Germany.
  • Publish analysis results in Open Access scholary contribution channels.

Research outcomes

Please find corresponding outcomes in the appendix below.

  1. [Kra2020] Kratzke, Nane. Redundancy Reduction in Twitter Event Streams, in ResearchGate, 2020, preprint ResearchGate Bibtex
  2. Details


  • Twista: Twitter stream recording and analysis command line tool (DOI: 10.5281/zenodo.845856)
  • Dataset: Monthly Samples of German Tweets (DOI: 10.5281/zenodo.2783954)
  • Dataset: The #BTW17 dataset (DOI: 10.5281/zenodo.835735)
  • Presentation: Der Bundestagswahlkampf 2017 auf Twitter - War der Ausgang abzusehen? (CoSA Seminar, 23 Oct. 2017)