Twitter network analysis with Gephi

Earlier this year I had a chance to do some social media network analysis and visualization with Gephi. This was done by scraping data from Twitter with NodeXL. Due the limitations of Twitter API, scraping enough tweets takes a lot of time. For this earlier project we used a Finnish tech company called Reaktor and their Twitter hashtag #ReaktorSpace. Reaktor had just launched its own space program, which caused some buzz in the Twittersphere. Back then with almost 700 tweets our study group managed to create the following representation.

reaktor_clusters

This, like any other graph it is an ordered pair G = (V, E) comprising a set V of vertices or nodes together with a set E of edges or lines, which are 2-element subsets of V (i.e. an edge is associated with two vertices, and that association takes the form of the unordered pair comprising those two vertices). So in this case Twitter handles (like ReaktorNow) are nodes, which are linked to other handles with edges.

With Gephi, we can calculate modularity, which helps us to identify communities within the network. This done by detecting the nodes that are more densely connected together than to the rest of the network and then coloring them. Nodes were sized by their average weighted degree, so the nodes with more connections with other nodes (in other words, Twitter accounts with more interaction) are larger. Finally, nodes were laid out with Force Atlas 2 algorithm to disperse groups and give space around larger nodes.

reaktor_clusters2

In this case, we identified communities as:

  1. The Employees and fans of Reaktor (pink)

Tweets from the company (@ReaktorNow) are retweeted by the employees and other followers/fans of the company

  1. Members of ‘academy’ world (green)

The space project was promoted in an event (#SpaceOnStage) of Aalto University. The director of European Space Agency (ESA), Jan Woerner (@janwoerner) was also present in the event and retweeted about the matter.

  1. Business people (blue)

Consultants of Reaktor were tweeting about #reaktorspace, as well as @masapeura, who is a director of new business in OP (one of the largest financial companies in Finland)

  1. Communications & PR people (orange)

Marketers of Reaktor (@jani_jaakko, @satutellervo) and other media / social media enthusiasts


Our conclusion was that the members of Reaktor family (cluster 1) are acting as an echo chamber for tweets, but not truly engaging with them. Reaktor managed to grasp the interest of some influential accounts (ESA with +500k followers), which could help to reach out more people who are interested in space, but don’t know about Reaktor. Some members of the business community were engaged, but this sector could be done better – for example explaining the business benefits and use cases of space tech in business.

References:

http://stackoverflow.com/questions/21814235/how-can-modularity-help-in-network-analysis

http://www.martingrandjean.ch/gephi-introduction/

https://nodexl.codeplex.com/

https://www.reaktor.com/expertise/space/

Leave a Reply

Your email address will not be published. Required fields are marked *