Study of a social network in gaming

Gaming groups, like any community of interest, experience a consistent background level of churn in membership. This can be due to life changes, shifting priorities, or evolving preferences for different game genres - players grow up, grow apart or move on.

Here we're looking at the changes in one such group, measured using the amount of time the members spend in activities with each other. The data is pulled from the Destiny 2 API covered in this article, taking a target set of player IDs and collecting instances where those individuals have been in an activity that included at least one other player in the same target set.

It's possible to see the stronger links between some individuals, shown by the thicker lines connecting the nodes in question. This tells us that these players have joined activities together several times over the given period. The weaker connections, the very thin lines between some nodes, suggest the two players may seldom cross paths even though they belong to the same clan.

This is expected. As in all friendship groups, sub-groups form - players drift towards those who have similar perspectives, interests or schedules and more often can be found in activites with this new subset. And while fractures within the group may seem at the time to be a complete surprise, with hindsight they're easy to see developing.

The buttons here let you filter for different periods in the life of the group, seeing how it evolves.

  1. The original small friend group (Cluster 1, in green), playing together regularly.
  2. Recruitment of new members.
  3. New members invite their friends to join.
  4. A distinct new group (Cluster 2, in red) forms led by pre-existing friendships.
  5. The split - a player is kicked from the group, and Cluster 2 quickly follows.
  6. Regrouping; still a little interaction between Clusters 1 and 2, but very reduced.
  7. Present day - no interaction at all between the two clusters, and the Cluster 2 is seeing fractures of its own.

This is a common and recurring cycle, and particularly as gaming groups increase in size - like any community, it requires nurturing and management. Could the group have seen this coming? Only by monitoring in-game interactions, as shown in the graph here. Should a group leader be monitoring in this way? I would say, 'no' ... it's an interesting study after the fact, but uncomfortably invasive at any other time. All of this data has been anonymised for that reason - while members of the two clusters might recognise the patterns here, it's not my intention to to call people out.

It's only a game, after all!



About the author

Lucy works in education, open data and civic tech in the South West of England (and anywhere else she's needed).

Currently available for project work.

"People don't want data; they want answers."

(c) Jargonautical Ltd 2020. Branding and logo design by Shufflewing