Ozean Media Announces @AnnaforFlorida as Winner of The ‘@’ Award 

Rep. Anna V. Eskamani is Recognized as the Twitter Power User for the Florida House of Representatives

Alachua, FL – May 17, 2023– Ozean Media, a political affairs firm specializing in research, data, and media, announces @AnnaforFlorida (Representative Anna Eskamani) as the winner of The ‘@’ Award for Florida House of Representatives.

“Rep. Eskamani is the clear leader among her peers in the Florida House in the use of Twitter,” said Alex Patton, Managing Partner of Ozean Media. “She not only broadcasts her own messages, but also engages with others on the platform. If other elected officials are looking to up their Twitter game, Rep. Eskamani is a model worth studying.”

“It is no secret that significant political communication has been happening on Twitter, and Ozean Media is studying the various ways elected officials use or don’t use Twitter. The best in class rather than only broadcasting are interacting with people outside the political bubble and are using rich media such as images and videos”, concluded Patton.

“We knew studying Twitter at a time of massive changes on the platform would be risky, but we believe it was worth it. We believe we collected the largest sample of tweets from Florida Legislators to date, and we have gained a greater understanding of how elected officials are using Twitter,” said Ben Torpey, Ozean Media Consultant.

The ‘@’ Award will be presented at to Representative Anna Eskamani at her convenience in late May 2023. More information about the award, the final top 10 ranking, week-to-week standings, and the final report can be found at: Twitter Final Report 2023 FL Legislature

About Ozean Media

Ozean Media is a strategic partner in political affairs, providing consulting and public relations services to drive the movement of audiences toward a specific policy or cause. Ozean leverages research, data, message development, and media to achieve our clients’ goals. More information can be found at ozeanmedia.comhttps://ozeanmedia.com.

About The ‘@’ Award

The ‘@’ Award recognizes the top Twitter user among Florida’s House of Representatives and Senate based on a proprietary algorithm that weighs factors such as follower count, tweets, retweets, likes, and engagement. The award was originally scheduled to collect data until May 5, 2023, but Twitter’s API changes restricted access to data, so the award is based on data collected through Feb 26 – April 28, 2023.

Methodology

Given a specified time interval (March 5 – May 7 or Florida Session’ sine die whichever is later) and a list of Twitter handles, the algorithm assigns a tailored weighting to variables including tweets, retweets replies, follower count, following count, and effective reach. The initial ranking was done for a period of the week before Florida’s legislative session started or Feb 26 – March 4.

Unfortunately, the period for rankings ended on April 28 due to changes in Twitter’s API that drastically changed and reduced access to tweets and other data on the platform.

Therefore, final rankings were decided using the period of Feb 27 – April 28, 2023.
Our Twitter algorithm, while recognizing Follower Count and number of new tweets, also attempted to recognize different uses of the platform – to include other activities other than simply broadcasting tweets.

Key Findings

• Representative Anna Eskamani is the Power User of Twitter in the Florida House.
• Senator Shevrin “Shev” Jones is the Power User of Twitter in the Florida Senate.
• Power Users use Twitter differently – Power Users interact not only broadcast.

While the algorithm had to take follower count into consideration and raw number of new tweets, it also indicates a correlation between higher follower counts and different types of tweets.

Members that not only tweeted, but also replied and retweeted others, correlate with higher follower counts. It appears, those that have been able to grow a large audience are interacting with other users, not only posting and using the tool primarily as a broadcast medium.

Members that not only tweeted, but included media (photos, video, and GIFs), correlate with higher follower counts.

Final Ranking

Representative Anna Eskamani (D, FL 52) and Senator Shevrin “Shev” Jones (D, FL 34) are awarded The @ Award recognizing their “Power Use” of Twitter during the 2023 Regular Session of the Florida Legislature.

Both Representative Eskamani and Senator Jones lead the pack in Twitter usage from the beginning, and they never looked back. Other elected officials vied for the 2nd through 10th spots, but Eskamai and Jones were significantly ahead of their peers.

Content

Content varies with large blocs of “Thanks fellow member x for passing the/my bill through committee/chamber”, holiday greetings, and tweets on the abortion issue.  

Below are word clouds of the tweets of House and Senate members.

House Tweet Word Cloud

Senate Tweet Word Cloud

Tagging/Mentions

When elected officials @ someone, they are most likely to @ a fellow Legislator or elected official.

The second most likely entity is a Florida press entity.

A cursory review of the tweet data indicates precious few interactions with constituents.

Hashtags

House and Senate members used Twitter hashtags infrequently during the session. The most frequently used hashtag was “#latergram.”

Most prolific Tweeters

As one would expect, the algorithm rewarded elected officials that tweeted often.

In the House, Representative Eskamai was the top tweeter with 768 total tweets.

In the Senate, Senator Linda Stewart (D, FL 13 – final ranking #4) was the most prolific tweeter with 364 total tweets.

However, Senator Stewart’s relatively lower follower count prevented her from climbing higher in the rankings.

Largest Following

Representative Eskamani shines when it comes to follower count with 90,523 followers.

Senator Jones lead the way in the Senate with a follower count of 35,004.

To put these numbers in context, the highest follower count of any elected official in Florida is Senator Marco Rubio at 4,518,971, followed closely by Governor Ron DeSantis with 4,139,420.

However, Rep. Eskamani has more followers than half of Florida’s Congressional delegation, and Rep. Eskamani and Senator Jones have more followers than a majority of Florida’s statewide elected cabinet members.

Interactivity of Tweets (tweets by type)

Our algorithm rewarded elected officials for interacting on Twitter rather than simply broadcasting. The algorithm assigned weight for Replies and ReTweets.

This is one of the main reasons elected officials with smaller follower counts saw increases in their over ranking.

For example, Senator Linda Stewart (final ranking 4) had a follower count of 3,848; however, 46% of her tweets were replies or retweets. Representative Alex Andrade (final ranking 4) has a follower count of 5,717 however, over 62% of his tweets were replies or retweets.

Media Use

On average, 40% of tweets of the top 10 Power Users in each chamber contained media (photo, video, or animated GIF). Of the tweets that contained media, the vast majority (80%) were photos.

Platform Use

While the algorithm did not use ‘platform’ as ranking criteria, the information was collected.

Over 85% of the tweets sent from the top 10 Power Users were sent using Twitter for iPhone.

The remaining 15% were sent using the Twitter web app and Twitter for iPad.

Few rely on third-party apps like TweetDeck or Hootsuite. However, those not in the top 10 appear to use third-party apps at a higher rate.

While this needs more research, it is our theory, higher usage of third-party apps may be an indication that staff is managing social media accounts rather than the member. We observe an increase in the use of third-party platforms with “official press accounts” of Florida’s Congressional delegation.

Final Thoughts

It is of note tinged with sadness that Twitter is making its API cost prohibitive to researchers.   

Without underwriting, it is unlikely that additional studies can be conducted, and we think that is a shame.   

It was fun while it lasted, but our Florida Legislature Tweet Tracker is coming to an end.

With one week left in the regular session of Florida’s legislature, Ozean was notified that Twitter has made changes to its API that have drastically restricted third-party applications’ access to tweets and other data on the platform. In response, we are no longer able to collect the data needed for the ranking algorithm.

As of this morning, we are unable to continue to collect data from Twitter. As a result, we are finalizing the @Award today and will no longer be able to track tweets from Florida legislators.

We knew that entering this project, we risked being at the whims of Elon Musk, who acquired Twitter in October 2022.  We are sadden to see our fears realized. 

While we are disappointed that we can no longer track tweets from Florida legislators, we are grateful for the opportunity to have done so for the 2023 regular session.

We hope that Twitter will reconsider its decision to restrict third-party access to its data.

In the meantime, we will finalize and issue a final report in the coming weeks.

The results by precinct results have been released for Florida.  Precincts are the smallest unit of analysis that we have election results, and I find them interesting.

So what does the data tell us?

In this post, we will only explore the results for Governor’s race in 2022.

Messy Data

Even then, the data gets messy – there are 168 precincts that have no votes cast in them, most of them have no registered voters in them.  A few have 1 or 2 voters.   Who knows?  We will exclude them as outliers.

Then there are the even weirder precincts.  There are another 109 precincts with less then 10 total votes cast.    Again, who knows?  We will also exclude these from our analysis as also outliers.

Data Description

There are a total of 6007 precincts in Florida.  We have excluded 277 precincts with less than 10 total votes cast.

We are exploring 5,730 precincts in Florida.

Election Results – Data Check

A quick double-check against – official results.  The state shows Ron DeSantis winning with 59.4% of the vote and Crist with 40% of the vote.  The LPF candidate (Roos) received .2% of the vote, and a NPA (Gimenez) received .4%.

Comparing our precinct-level data file, it is almost an exact match.

Somehow Governor Ron DeSantis is down one vote, but we continue on.

When we exclude the previously mentioned precincts,  DeSantis loses 249 votes, Crist loses 148, Roos loses none, and Giminez loses 6.   In the 277 excluded precincts, we lose 403 total votes or less than .006 of the vote.  Immaterial.   

Final Data Set

The final dataset we are exploring is 5,730 precincts.  Those precincts cast 7,770,995 votes.

DOWNLOAD THE FILE

In this dataset, DeSantis won 69% of the precincts, Crist won 30.1%, in less than 1 % of the precincts, there was a dead tie.

FUN FACTS:

The largest precinct with a tie?  ALA056.  (Alachua 56 – Covenant Presbyterian Church).  In the books with a 66% turnout, 1,093 votes for DeSantis, 1,093 votes for Crist..

 

The county with the most precincts ending in a tie?  Palm Beach County – 5. 

Winners

DeSantis won 69% of the precincts or 3,963 precincts.

Crist won 31% of the precincts or 1,752 precincts.

On average, the spread between DeSantis and Crist was 33%.

In precincts won by DeSantis, the average spread was 33% (67%-33%).  In precincts won by Crist, the average spread was 31% (65%-34%).

Divided Florida

17% of Florida’s precincts have a delta of less than 10%.   Those precincts have a registration of 2,518,332 voters and 1,271,336 votes cast.  A 51% turnout.

82% of Florida’s precincts have a delta of more than 10%.  These precincts have a registration of 11,970,376 and 6,499,659 votes cast.  A 53% turnout.

Interestingly, the highest turnout by precinct is found in precincts that had a delta of between 30% and 50% – 57% of those voters turned out.   (I am fairly certain there is a research paper exploring that data point.)

Competitive Precincts

It was my intent to map the precincts that have a delta of less than 10%.  

Dr. McDonald of the University of Florida and the US Election project has the closest thing we have to a complete shapefile for the entire state of Florida. 

There is a glitch (at least in doing a quick analysis) in that some counties report data using a precinct name that differs from the shapefiles (DADE, BROWARD, PALM BEACH, OSCEOLA are the large ones.)   I started to remap them, and got Dade and Broward done, but Palm Beach may have renamed their precincts all together…so a couple of holes that I just don’t have the time to go into and match. 

However, we can look at most of I-4 (excluding Osceola & Lake).   

Conclusions

What can anyone draw from this cursory review?  

Just that few precincts in Florida are truly competitive.   

There is a debate in political science on whether this sorting is happening on purpose – with politics driving “the big sort.”

Bill Bishop wrote The Big Sort: Why the Clustering of Like-Minded American is Tearing Us Apart in 2008 “Armed with original and startling demographic data, he showed how Americans have been sorting themselves over the past three decades into homogeneous communities — not at the regional level, or the red-state/blue-state level, but at the micro level of city and neighborhood.” (thebigsort.com)

Of course, Morris Fiorina, has a different take.  “The Big Sort” That Wasn’t: A Skeptical Reexaminationwhich is summarized in The Myth of the “Big Sort” states that “claims about geographical sorting have always struck us as somewhat questionable.”

Fiorina, one of my favortie contrarions, critques Bishop for his reliance presidential election returns saying they are often inconsistent for other offices.  Fiorina uses county level data to refute Bishop.  

Fiorina sums up his critque by saying ” There is no evidence that a geographic partisan “big sort” like that described by Bishop is ongoing, and even if it were, its effects would be far less important than Bishop and those who support his thesis fear.”

Yeah, about that…

17% of the precincts in Florida are within 10 points in a Governor’s race.  17%!

Maybe time to re-fresh the research…and while we are at – let’s explore that bump in turnout between 30%-50%.