INFORMS NEWS: A Twitter recap of INFORMS 2015

By Chris Marks and Tauhid Zaman

Figure 1: The #INFORMS2015 status update with the most retweets.

Figure 1: The #INFORMS2015 status update with the most retweets.

Following the 2015 INFORMS Annual Meeting in Philadelphia in November, we collected Twitter status updates, or “tweets,” using the Twitter search API and the python methods from the INFORMS Twitter Data Mining Tutorial (available at http://web.mit.edu/cemarks/www/tutorial.html). The search query we used was “#INFORMS2015,” which returned 2,097 tweets spread out over approximately a one-week period encompassing the annual meeting. Of these status updates, 878 were “retweets,” where a user re-posted a status update originally posted by a different user. Figure 1 shows a screen shot of the most retweeted tweet, which is from user @lauramclay, and was retweeted 37 times.

The pulse of INFORMS 2015: Figure 2 depicts the tweet rate, in tweets per hour, of the “#INFORMS2015” hashtag. We observe three spikes in this tweet rate, each occurring at approximately the same time of day on Saturday, Sunday and Monday, as indicated in the plot.

A summary review of the content in each of the spikes reveals interesting information about the conference. The first spike, occurring on Saturday, contains primarily of tweets about the start of the annual INFORMS meeting. The Sunday and Monday spikes consist of many tweets from a single user, @michaelhoffman, who gives updates and quotes from plenary talks and other presenters during these periods.

The language of INFORMS 2015: To get an idea of what INFORMS 2015 was about, at least according to people who use Twitter, we created the word cloud in Figure 3 using the “wordcloud” package in Python. This figure depicts the most commonly occurring words in the “#INFORMS2015” tweets, where the size of each word increases with the frequency of occurrence in the data. As can be seen, “analytics” and “data” are the hot topics of discussion on Twitter for INFORMS.

Figure 2: Tweet rate of “#INFORMS2015” during the 2015 INFORMS Annual Meeting.

Figure 2: Tweet rate of “#INFORMS2015” during the 2015 INFORMS Annual Meeting.

Figure 3: Word cloud of “#INFORMS2015” tweet content.

Figure 3: Word cloud of “#INFORMS2015” tweet content.

The connections of INFORMS 2015: Finally, we provide a network depiction of tweet “user-mentions” in the “#INFORMS2015” data. The plot in Figure 4 displays Twitter users as nodes in a directed graph. An edge exists from User A to User B if User A mentioned User B in an “#INFORMS2015” tweet. Table 1 provides details on the top-mentioned accounts in the “#INFORMS2015” data set. Not surprisingly, @INFORMS and @INFORMS2015 dominate the list. The other accounts belong to users who gave talks or presentations at the conference.

Figure 4. User-mention graph from “#INFORMS2015” original tweets.

Figure 4. User-mention graph from “#INFORMS2015” original tweets.

USER NUMBER OF MENTIONS NAME DESCRIPTION NUMBER OF FOLLOWERS NUMBER OF FRIENDS
@INFORMS     30 INFORMS INFORMS is the largest professional society in the world for professionals in the fields of analytics and OR/MS 4967 453
@INFORMS2015 25 INFORMS2015 The INFORMS Annual Meeting, with more than 1K sessions & just over 5K papers, covers broad landscape of O.R. research & practice. Use hashtag #informs2015. 424 5
@agrobins 10 Anne G. Robinson Advanced analytics enthusiast focused on strategy, supply chain & social media. Newfoundlander. Past president of INFORMS (http://t.co/KCPClAO7z3) 703 315
@thserra 9 Thiago Serra PhD candidate at CMU, optimization enthusiast, and proud father of one 317 346
@miketrick 9 Michael Trick Professor of Operations Research, Sports Scheduler, Blogger 2416 485
Table 1: Twitter account descriptions of top-mentioned “#INFORMS2015” users.