#NetNeutrality hashtag visualizations
I captured two hashtags related to net neutrality trends on Twitter. Here are the progressions for #NetNeutrality and #SaveNetNeutrality with analysis of some anomalies I found in each hashtag.
#NetNeutrality
Date Range: Nov 21 (4:30pm est)— Nov 21 (9:30pm est)
Tweets: 4,499
Nodes: 5,489
Edges: 6001
Communities: 600
Impressions: 9,307,308
When I analyze hashtags, I’m looking for unusual patterns in the graphs. Applying filters in gephi removes the noise of the network in order to see any anomalies that might exist.
I filtered the #NetNeutrality network by edge weight, removing all edges with a weight of 1. This removes all edges that represent a single interaction between two nodes. For example, if I retweeted one of @MarkRuffalo’s #NetNeutrality tweets, the action would create a single edge between my account and Mark Ruffalo. Applying this filter would remove the single edge between our accounts. Removing those single interactions leaves only edges between nodes that represent multiple connections. These are accounts that are tweeting multiple #NetNeutrality tweets.
Multiple tweets could be a sign of automation (tweeting and/or retweeting more than normal) or it could also be a sign of an account that cares about a cause and is tweeting a lot about the topic, which is what I observed in the #NetNeutrality tweets.
Typically I will zoom in to look at these heavy edges to see what the accounts look like, and often find the nodes at the ends of these heavy edges are automated. This is not what I found in the #NetNeutrality tweets I looked at.
For example the 5 accounts highlighted below connecting to Ajit Pai all appear to be normal accounts tweeting under 10 times per day. Each account retweeted multiple tweets that mentioned @AjitPaiFCC’s account which is why the edges connecting their accounts to his are heavier. The only semi-unusual account was @ComputrRage, which according to its bio is a “Social Media/Virtual Assistant” that tweets horoscopes and occasional news links. This account is tweeting from an app called Twittascope but only averages around 2–3 tweets per day. @ComputrRage retweeted several tweets about net neutrality that mentioned Ajit Pai’s account which is why it has a heavy edge connecting it to Ajit Pai.
I looked through many of the nodes with heavier edges and found the same results for all of the accounts I checked. They’re not automated accounts, they all appear to be normal accounts that tweeted a lot about net neutrality.
In the below graph zooming in on @billprady’s account, @AliSWasTaken and @InsectOfJustice tweet about 9 and 13 times per day respectively. They are active accounts and they both tweeted multiple tweets with the hashtag #NetNeutrality, but they seem like normal accounts.
#NetNeutrality
Date Range: Nov 21 (4:30pm est) — Nov 22 (10:15am est)
Tweets: 14,566
Nodes: 13,915
Edges: 15,226
Communities: 1,494
Impressions: 26,412,495
#NetNeutrality
Date Range: Nov 21 (4:30pm est) — Nov 24 (10:00am est)
Tweets: 42,450
Nodes: 38,684
Edges: 44,988
Communities: 3,999
Impressions: 80,217,537
I filtered the final graph of 42,450 #NetNeutrality tweets by edge weight again and found many accounts tweeting up a storm about net neutrality. It’s an important example in this type of analysis, just because something looks like it might be unusual doesn’t mean it is, which is why it’s necessary to reference the individual tweets for context before jumping to conclusions.
This is a large, complex network of people connecting multiple times to the FCC and Ajit Pai and several other main accounts. The orange area in the upper right are telecom companies: Verizon, AT&T and Comcast. The light blue above the FCC is Fight for the Future, a non-profit organization that supports freedom of expression, privacy online and maintaining net neutrality. The mint green formation in the bottom left area is the ACLU National account.
The heavy pink edge (near netflix) grabbed my attention. When I zoomed in to that area, I found a 20 year old YouTuber from the UK named Helen.
@imjusthelen tweeted a thread about Net Neutrality, she tagged several accounts in her thread including Twitter, wetheunicorns, YouTube, facebook, Google, AjitPaiFCC, BrendanCarrFCC and mikeofcc and she put the hashtag #NetNeutrality in every tweet.
I looked at several other accounts like this and found similar results as before: accounts that appear to be normal people who tweet at a normal human rate and support #NetNeutrality. Many people tweeted multiple times about net neutrality and their activity created a huge web of connections, most of which appears to be organic.
The next two graphs are filtered by degree range to remove smaller nodes and focus on the influencers. Just above Ajit Pai’s account there’s a node that stands out, almost like a vortex connecting many different nodes in the network.
Zooming in to the odd node you can see it’s an account called @ham_gretsky.
@ham_gretsky is using an app called Retweeter_ to retweet net neutrality tweets, about 61 tweets per day. Judging by its bio and activity, it seems like @ham_gretsky is using some automation to aggregate and boost #NetNeutrality tweets. It might be considered a cyborg (a human using automation to tweet). I tried to contact this account to ask about it’s activity. If it replies I’ll update this post.
As a supporter of net neutrality, I don’t really have a problem with a bot or a cyborg that boosts #NetNeutrality tweets. As long as an account is properly labeled and is not masquerading as a real person, I don’t think this kind of automated activity is malicious. Maybe someone who wants to repeal net neutrality might feel differently about it, but the account is not being abusive or violating Twitter’s terms of service.
I captured tweets for #SaveNetNeutrality and visualized the dataset at various intervals from November 21 through November 25 to watch the progression.
#SaveNetNeutrality
Date Range: Nov 21 (11:00pm est) — Nov 22 (10:30am est)
Tweets: 6,863
Nodes: 6,367
Edges: 6,970
Communities: 1,058
Impressions: 11,819,324
#SaveNetNeutrality
Date Range: Nov 21 (11:00pm est) — Nov 24 (12:00pm est)
Tweets: 17,073
Nodes: 14,573
Edges: 18,128
Communities: 2,279
Impressions: 35,179,455
The node at the end of the heavy, purple edge connecting with Ajit Pai’s account stands out. That’s ham_gretsky again, the account I mentioned above that’s using an app called Retweeter_ to aggregate and retweet #NetNeutrality and #SaveNetNeutrality tweets.
@sketchpads19’s account seemed like it might be using some kind of automation and I could see it was tweeting a lot of net neutrality tweets. I thought it might be a cyborg, but I was wrong.
I chatted with @sketchpads19 and learned he’s a human who works in IT. He told me he tweeted so much about net neutrality that Twitter thought he was a bot and kicked him out of his account twice and made him verify his identity to log back in. I asked him why was he tweeting so much about net neutrality and here’s what he told me:
When it comes to freedom of information, freedom to express ones opinion, and freedom to choose where you go online, Net Neutrality is something that is extremely vital. By removing Net Neutrality, companies can slowdown, hinder, and possibly restrict what sites you go to. Whether those sites be news outlets, blogs, YouTube, or even Twitter, all behind the guise of paying more for online content. What is to stop an ISP from blocking sites or restricting site that show negatively about that ISP? It’s happened in the past, and they are trying to do it again. — Chris Ivanski (@sketchpads19)
The activity I found in #SaveNetNeutrality was similar to #NetNeutrality; many real people who support net neutrality and are tweeting A LOT about it.
I looked for automation in both hashtags but what I found in the datasets I analyzed was negligible. The following infographic from Tweet Archivist shows the sources of #SaveNetNeutrality tweets.
The most common devices used to tweet around the world are usually Android, iPhone and Web Client. Twitter Lite and iPad are not unusual and here you can see there are a few automation programs being used in #SaveNetNeutrality.
Retweeter_ is the app being used by @ham_gretsky who I highlighted above and was the source of 579 #SaveNetNeutrality tweets or 0.034% of this sample. TweetDeck is a common tool used by community managers that supports multiple accounts. 106 tweets out of a total 17,073 tweets, or 0.006% of the #SaveNetNeutrality tweets in this sample, were sent using TweetDeck. Tweetbot for iOS is another app that supports multiple accounts like TweetDeck and provides statistics on Twitter activity (follows/unfollows). Overall the amount of automation in this hashtag was small.
This infographic also shows the number of impressions from Tweet Archivist, a statistic I’ve cited throughout this post. Tweet Archivist has the following information on their website about how they calculate impressions:
Impressions are the total number of times that the tweets of an archive have been delivered to Twitter streams. Of course, not everyone who receives a tweet will read it. As such, impressions are the largest possible audience for the given archive. Paid advertising works similarly; even though an ad was displayed on a website, there is no guarantee that a person actually saw it. Also, note that impressions does not deduplicate users, so if the same person sees a given hashtag twice, it counts as two impressions. Note that, because replies are only delivered to common followers’ timelines, they are calculated as a single impression.
#SaveNetNeutrality
Date Range: Nov 21 (11:00pm est) — Nov 25 (11:00pm est)
Tweets: 20,815
Nodes: 16,803
Edges: 21,545
Communities: 2,387
Impressions: 44,256,310
Initially when I saw some unusual patterns in these two hashtags, I assumed I was going to find automation. I was surprised to find most of the accounts that seemed like they might be artificially amplifying their tweets were just normal accounts that were tweeting more than usual because they care about net neutrality.
I also captured #NetNeutrality tweets in May 2017. I didn’t do an in depth analysis of those tweets but I uploaded the graphics to the Internet Archive if you’re interested to see some historical data. There was an account in May called @NetFreedomBot — similar to @ham_gretsky that aggregated and retweeted #NetNeutrality tweets but that account was suspended.
Two hashtags #NetNeutrality and #SaveNetNeutrality made a combined total of 124,473,847 impressions from November 21 to November 25 and I assume the notifications for @FCC and @AjitPaiFCC are lit. Most of the activity I looked at appears organic.
If you haven’t read the natural language processing analysis by Jeff Kao that found 1.3 million fake anti #NetNeutrality comments, I highly recommend reading that next.
