We are just getting back from this years’ HASTAC conference in Lima, Peru and it’s time to take stock of what we discussed during the conference. The presentations covered an extraordinary variety of topics associated with the internet, learning, history, games, humanities, and digital media. Inspired by Ernesto Priego’s Twitter archive of the conference, and putting to good use the Twitter archive we created for the conference, we decided to mine this data to see what story the Twitter stream tells about the HASTAC 2014 in Lima, Peru. The dataset includes 3919 messages tweeted in an undulating behavior but with prominent peaks on April 24-27. The peaks last just the mornings and afternoons which coincide with the period of the conference.
The first message was tweeted from an iPhone on Monday, Apr 14 at 3:02 am by user cliffmanning. The message reads “If anyone is looking for cheap accommodation for #hastac2014 Dimar Inn is friendly and almost next door casa andina http://t.co/km2k1aqrER.” The last message in the dataset is a retweet from @waltercurioso by user @Reciaria using an Android device on Thursday, May 01 at 2:07 pm. The tweet is in Spanish and reads “Repositorio #ALICIA @ConcytecPeru busca generar cultura de intercambio información @ocram @EnContactoATV http://t.co/Pmn. [Repository #ALICIA creates a culture of information sharing],” which is a call to the National Digital Repository of Pery. The four most retweeted messages were tweeted by Twitter account @HASTAC, @NazcaTheMad, @CathyNDavidson, and @jshhnn. Below is the timeline of the most retweeted messages.
As indicated by the y-axis of the figure above, even the most retweeted messages include around 9 replications only. The number of messages varies per day, as indicated by the x-axis, with a minimum of 9 and maximum of 13. This is not a lot of retweets, but it goes to show that a particular set of messages have a stronger appeal to the public attending the conference compared to messages that were not retweeted. They also provide a good overview of the topics discussed during the conference. @HASTAC tweet is the most replicated message with 13 retweets discussing the role of content in connected learning. @NazcaTheMad points out that digital humanities and openness are buzzwords co-opted by the market, and @lizlosh message mentions the very high retention rate of the network of scholars and artists FemTechNet. Finally, @CathyNDavidson tweets points out to Will Ross post about Net Neutrality and to a summary of the conference.
The tweeted messages cover a wide range of topics. To make sense of the data, we modeled and grouped the tweets into three central topics using the variational expectation-maximization (VEM) algorithm. The first topic revolves around conference panels and presentations on digital humanities, narratives, and collaboration. The second topic is mostly devoted to the keynote speaker Mitchell Baker and the HASTAC community. The third and last topic focuses on conference logistics, with terms like conference program, people, public videos, and workshops. The first topic also grouped the discussion about Net Neutrality, which was a tweeted by Cathy Davidson, discussed by Mitchell Baker, and that I live blogged during the conference. The third and last topic also grouped messages on connected learning, education, and pedagogy. The word cloud below is coded in green, red, and blue to match the abovementioned topics.
These topics can also be inferred by looking at the most tweeted URL links during the conference. The two most frequent links are associated with conference logistics. The first directs users to the conference live stream and the second takes users to the conference program. The third and fourth most frequent links are associated with conference presentations. One points to Julie Keane’s presentation on Open Badges and the other to Cathy Davidson’s call for Mitchell Baker’s live-streamed keynote talk in the conference. Another URL that was very tweeted points to the presentation of Josh Honn about the Czech-Brazilian writer and philosopher Vilém Flusser. The tweet with a link to the presentation was retweeted over 10 times.
We also explored the sub-networks created by users by resorting to retweets, @-mentions, and co-occurrence of hashtags. These networks are highly clustered with a few tweets being retweeted retweeted over and over again, a few users being particularly active at exchanging messages, and one or two hashtags grouping most of the hashtagged keywords used during the conference . Except for the @-mention network, which presents far more structural holes, the remaining networks are densely connected to a core of users or hashtags that boosted the conversation. This is shown in the graphs below, with retweet and hashtag networks connecting to dark hairball at the center of the image while the @-mention network presents a more complex structure.
If you’re interested, you can also check a more detailed (and surely prettier) visualization of the networks. Here’s the hashtag network and here’s the @-mention and the retweet network, with retweet edges code-colored blue and @-mention edges code-colored red. The latter graph shows that the network of users exchanging messages does not overlap very much with the network of retweets. This is not surprising as users exchange information during the conference with selected peers, but only retweet messages that are relevant to the broader community. A snapshot of these overlapping networks is also available here.
This was my first HASTAC Conference (of many, hopefully) and I did have an amazing time connecting with HASTAC scholars and figuring out where humanities and arts are heading to in our all-digital mediascape. I saw some downright fascinating presentations and I’m very happy researchers in the humanities area have such a great venue for discussing and presenting ongoing work.
We are now getting ready for the HASTAC NSF EAGER Conference at the PhD Lab in Digital Knowledge at Duke University. This one day-long workshop will be held on May 28 on the provocative theme: “Can Analysis of Big (and Sometimes Messy) Data Facilitate Collaboration? Methods, Models, Tools, Best Practices, and Next Steps in Exploring Multi-Institutional Interdisciplinary Collaboration Online and Onsite.” If you’re interested in this discussion, make sure to register via Eventbrite.