The Last Three Tweets: Social Media and Election Campaigns

The increasing availability of data is pushing the boundaries of what was once imagined possible in public diplomacy. Data science has the potential to draw large data sets into the study and practice of diplomacy, and allow diplomats and scholars to become comfortable engaging with and analyzing increasingly large and often unstructured data. As Google Chief Economist Hal Varian claimed, “the ability to take data—to be able to understand it, to process it, to extract value from it, to visualize it, to communicate it—that’s going to be a hugely important skill in the next decades.”

However, while processing power and tools to analyze data have expanded rapidly so that users can now search large amounts of data quickly and efficiently, it is still up to both subject matter experts to interpret the results and public diplomats to bridge the last three feet and deliver results.

One area in which data science techniques can provide insight for public diplomats is the use of social media around election campaigns. Elections often produce large quantities of data, and “election observation” has previously been linked with public diplomacy in thematic areas such as democracy and human rights. The growth of social media platforms creates greater opportunities to observe, report, and expose election violence and misconduct.

Election observations also create an opportunity for diplomats to recognize influencers and understand the specific interests of particular communities identifiable online--insights which are particularly valuable to the processes of diplomacy and public diplomacy after the election.

The 2015 election campaign in the UK highlights two potential opportunties for identifying influencers: first, analyses of the way individuals interact with political parties and candidates via social media, and second how conversations about the election take place using specific terms or hashtags. In both cases important insight can be gained from analyzing which messages are particularly popular, and the extent to which political conversations remain within specific communities. 

Option 1: Track the interaction via social media with political parties and candidates

I recently worked on an analysis of the interaction with and about political parties and candidates via Twitter, with UK-based think tank Demos. An analysis of 1.5 million tweets by or about candidates from early in the campaign examined the way users shared information via retweets. It showed that users tended to group together in clusters which are, broadly, indicative of parties, countries and shared ideologies. As shown in the image below, some politicians are particularly prominent within each group (often party leaders, such as David Cameron or Ed Miliband), but in other groups users commenting on candidates are prominent (for example blogger Guido Fawkes or Scottish comedian Frankie Boyle). See image above for a visualization.

Building on this analysis, a larger version using five million tweets was produced for the BBC TV show Newsnight. Using the same criteria over a longer time period, this version showed that despite the ebb and flow of the respective campaigns users remained largely within their respective party allegiance. The analysis, broadcast by the BBC, shows the groups in greater detail and highlights that the most popular tweets tended to remain within a specific political group. The full network can be examined here and a longer explanation of the analysis is in this BBC video:

Option 2: Track conversations about the election through hashtags

An alternative approach, produced in collaboration with FourTold, focuses on conversations around specific hashtags relating to the UK election – for example, #GE2015 and #UKGeneralElection. This approach has the advantage of enabling social media users to be influential, if they have something to say which others value. Users emerge as influential because what they say resonates with other users, rather than their status within a political party. This approach can be particularly useful in understanding the views of those either disillusioned by or excluded from the political process – for example, in analyzing the protests following the 2009 Iranian Presidential Election. This is because this approach includes users who have things to say about policy or events, but do not necessarily engage with politicians or political parties.

Using this approach to analyze the UK election campaign, more commentators, comedians, and satirists are included, providing a wider perspective on UK politics. 

Equally, understanding this data can provide insight into where social media users got the information which they valued sufficiently to share, and the types of devices they use. For example, analysis of the current discussion of the UK election on Twitter shows that this is the first UK General election where the tweets are predominately sent via smartphone, rather than desktop, a shift which becomes even more pronounced if tablets are included as smartphones rather than as their own catagory. This is an important insight for diplomats planning future activities as they seek to interact with politically engaged social media users; it is after all vital to use the appropriate medium to bridge the last three feet.

Each approach provides different perspectives on the 2015 UK election campaign. The first focuses specifically on key members of the electoral process. The second provides a broader view of discussions related to the election, but neither assumes politicians are the most important voices in this discussion, nor even necessarily includes all the content from politicians (as they may not use the tags in their tweets). Instead, it provides a view of the voices which resonate with other users. After the election, the information gleaned from the data produced by both approaches would allow diplomats to recognize users who were influential during the campaign and who are likely to be influential and worth engaging during subsequent public diplomacy activities.

Ideally, the era of big data will not drown diplomats in a sea of information, but will yield useful ways to derive insight and develop innovative strategies from that data. For example, those seeking to follow Open Source or Collaborative Public Diplomacy strategy can use the approach to find communities of potential allies.