It is now common in the field of digital humanities: the public of the lectures is at least as much present on Twitter than physically in the room. From July 7 to 12, the annual international conference of Digital Humanities DH2014 was held on the campus of Lausanne, at the invitation of LADHUL (UNIL) and DHLAB (EPFL) and brought together more than 750 researchers from all around the world. During this same period, 16,000 tweets were sent by almost 2000 users. This post, as well as the one from my colleague Yannick Rochat, provide a few visualizations of these intense discussions to the community.

The complete #DH2014 Twitter network

The graph below represents all the “mentions” contained in the #DH2014 tweets (a tie connects two users when one mentions the other at least once in a message). The size of the circles indicates the number of tweets sent. The intensity of the color depends on the number of incoming mentions (in degree): the more a user is mentioned, the clearer the color (from blue to white between 0-100 mentions, and white for more).

It appears that the most active people are not necessarily the most mentioned people. When a member of the public “live-tweet” a conference, the followers do not respond to all tweets. Inversely, the keynote speakers are massively mentioned during their talks (Ray Siemens, Melissa Terras and Bethany Nowviskie appear as small white circles above).

This network is very dense (even with a strong force-directed algorithm). It is interesting to observe that no community is separated, even partially, from the main cluster: probable sign that users follow the hashtag with discipline and responded to everyone rather than follow some particular individuals. In detail, however, the connoisseur recognizes interest or language communities.

In the right pane, we observe that the list of users who have sent more than 100 tweets does not correspond with that of users that have been mentioned more than 70 times. In addition to the three keynote speakers mentioned above, institutional accounts like @DH2014Lausanne, @ADHOrg, @PelagiosProject and @HumanisticaDH were also much mentioned.

#DH2014 day by day

The mass of tweets is unevenly distributed in time, something that the complete graph does not reveal. Yannick Rochat has produced a histogram that shows the number of tweets sent every 10 minutes. It easily distinguished keynotes (especially Bruno Latour on Tuesday) and conference sessions. Coffee breaks and poster sessions are clearly less suitable for tweets.

In a more visual and instinctive approach, global signatures of the six daily graphs are also instructive:

If you have published a report on the conference, please leave the link below in the comments, I’ll add it to this post!

Tool: Gephi / Dataset: Y.Rochat