Brexit debate on Twitter Week Ending March 27th

During the period 21st March to 27th March 2016 we collected approximately 300k Tweets relating to the “Brexit” referendum in the UK.

Update: The O’Reilly book “Graph Algorithms on Apache Spark and Neo4j Book is now available as free ebook download, from

Here is a summary of our analysis of the most influential Twitter Users during the period together with some commentary on interesting observations.

Top Influencers

Network Map

In a previous analysis we noticed that the two main groups were a strongly united Leave group and a group consisting of Remain and a wider set of mainstream media Users.

Here we see the same split which can be easily observed from the network map.

The colours in the map represent the communities that are detected by our community detection software. At a high level the colours show the two main groups.

However, communities are detected which consist of smaller groups of Users who share more narrow interests.

Here the colours are representing a more detailed view of the sub groups within the main communities.

This provides further evidence that the Leave campaign is a more tightly knit group with less diverse sub groups than the Remain campaign.

Polly Toynbee

Here is what we can quickly see from our dashboard on the Tribe that grouped around Polly Toynbee.

Here is a list of more of the top Tweets.

And here is the article that Polly Toynbee published.

Katie Hopkins

Katie Hopkins also made a Tribe and this time we can get an indication of the main topics of interest from the Topics section.

The top Tweets in this Tribe

One of the Tweets Katie Hopkins sent which had a large impact.

Allison Pearson

Allison also formed a community and was also focused on the security issue here is an example Tweet.

Different Styles

We noticed something quite interesting about how the three women achieved their influence.

Here are the three Ego Networks for each of them. An Ego Network is a network which shows all the other Users in the network that are just one jump away from them.

Polly Toynbee

This is a small network with one other Influential User in the network “The Guardian” which has 5m followers. The influence of Polly Toynbee would seem to be inextricably linked to the Guardian Newspaper.

Katie Hopkins

This is a much bigger network. The second most visible user in this network is David Vance who has 20k followers. In Katie Hopkins case this look very much more like a popular following of ‘ordinary’ people.

Allison Pearson

This network is also quite big and has one other main User in it — Nigel Farage who has 260k followers and is a key player in this debate. The conclusion you might draw from this is that Allison Pearson is an ally of Nigel Farage.


Everything is Obvious *Once You Know the Answer.

The fact that Polly Toynbee is an influential journalist working for the Guardian does not seem like much of a revelation if you are familiar with the domain of UK political journalism.

However, this analysis was done from a starting point of zero knowledge of Polly and her role in this debate. The analysis is done mostly through machine learning with the network maps and dashboard tool illustrated provided to a very inexpert analyst — me!

I hope this illustrates the power of the tools and how information can be uncovered about influencers and the stories they are influencing.

For an expert analyst of the UK Brexit campaigns there are much deeper insights which can be gained rapidly using these tools.

For more information about the methodology see this presentation which I prepared for Oxford University.

Social Network Analysis Presentation
Free download: O’Reilly “Graph Algorithms on Apache Spark and Neo4j”