Harris Surges with Announcement
This week Kamala Harris kicked off her 2020 campaign in earnest. Much like Elizabeth Warren (pink) when she announced her exploratory committee, Kamala Harris (blue silver) had consistently high mention counts following her announcement. Their media and campaign events produced sustained online conversation.
Interestingly, the surges for Warren, Harris and also Bernie (blue-violet) have a similar shape: a peak followed by several days of sustained conversation, degrading over time. It will be fun to see if this is a common feature of successful media outings. Contrast that with the stark presence of a Trump tweet about Elizabeth Warren earlier this month, or the less sustained debut of Castro (olive green).
While the main event remains Harris versus Warren, the other candidates are much more evenly represented in this week’s counts.
The top ten candidates all had mention counts in the “Tier 3” range from last week’s post.
But Harris’ count of 244K+ mentions extends an already substantial lead from last week. And many of these “candidates” might not run.
Scaling Down the Bias Plots
We made some adjustments to our Bias Distribution Plot this past week. The results help illustrate more about how the scoring system works and also how it can be read to extract insights about the conversational terrain.
First, we reduced the default source count from 1000 users to the top 100.1 This was a measure to improve load time. And we will rarely show only 100 users in plots for our posts. But the default weekly distribution this week shows some results worth mentioning.
First, the number of times (some) people tweet about candidates in a given week, as indicated by node size, is increasing pretty fast. Second, the most actively engaged Twitter users in the Democratic primary conversation have an apparent left-of-center bias.
And Scaling Them Up
Of course, we can also show plots with more than the previous count of 1000. This is especially useful for daily and per-candidate views of the data.
When we were finally able to look at the top 10,000 users this week, we could clearly identify a pair of clusters. These left and right clusters are entirely expected. Unexpected is that they have been slowly drifting toward the yellow cluster of unscored users at the bottom.
See my post two weeks ago for more on how these user scores are computed. Suffice it to say, there are two methods for scoring a Twitter user: 1) by the links they tweet or; 2) by users they retweet or reply to. We only use (2) when we don’t have (1). The unscored users in the yellow cluster are interacted with enough by the members of the red and blue clusters to pull both of these scored partisan clusters toward the unscored one, centered at (0,-2). 2
The Four Bs
Chuck Todd, among others, has clumped together an artificial group of candidates: those whose names begin with B. The implication is that each of these candidates is having a moment.
Certainly each of Booker, Biden, Beto and Bernie are mentioned at much higher rates this week than last. Each held campaign and media events this week. For Booker, Biden and Bernie, these were highly conventional appearances. And for each, a boost in attention was indeed the result. For Beto, whose “media events” this week were much less conventional, the reception was somewhat more negative.
Bernie is in the midst of a sustained mention surge. It is much smaller than, but concurrent with, Kamala Harris’. Like Harris’, Bernie’s surge follows media appearances on Martin Luther King Day. While it’s not yet clear he will run, it’s worth taking a closer look at the internals of the Bernie surge. He is almost always mentioned by himself.
And his distribution is full spectrum, with the largest cluster located left-of-center.
But the plot also shows a relatively large number of users tweeting links to right-wing sources.
The Fifth B: Pete Buttigieg
Pete Buttigieg announced his candidacy this week. On the first day he was mentioned in over 8000 tweets, almost exclusively by himself.
And on the second day he was mentioned in 4500 tweets.
On both days, Buttigieg’s mention distribution was dominated by 1) users new to our corpus3; and 2) users circulating left-of-center links.
Given the overwhelmingly positive tone of most of these tweets, Pete Buttigieg is positioned to do well when the mechanics of the race avail him of more “free media.”
Yet Another B
I’ll leave Michael Bennet for next week.
- Specifically the users who mentioned candidates in our corpus with a frequency in the top 1000 or 100, respectively, during the time period. Here it is a week.
- We fixed it by ignoring the “scores” of unrated users. All users are re-scored as they tweet, so the clusters are slowly drifting back to their “natural” locations.
- In other words, mentioning a candidate for the first time since November