Diving into Boston political Twitter

Twitter is the undisputed social media venue for national political conversations - presidents, senators, and reporters keep up a running stream of statements and reactions that compete with cable news as the background noise of Washington.

Is the same true for city politics? A smaller geography means that real life conversations are more possible, but a more connected political ecosystem might also make a social network like Twitter more vibrant. 

We set out to explore the political Twitter of Boston - its elected officials and candidates - in order to get a better sense of how the app is used, and who is connected with whom. 

For Boston electeds, we looked at City Councillors as well as State Representatives and State Senators whose districts include parts of the city. For candidates, we mapped the accounts of those currently running for Mayor of Boston or Council At-Large.

See the data of who elected Boston officials follow on Twitter here (more 8,225 accounts included) and keep reading for some analysis of our findings.

City Council uses Twitter more than the State House

While we can’t know how long elected officials spend scrolling Twitter throughout the day, it appears that it is more utilized in Boston than in the State House at large.

This year, Lisa Kashinsky took over Politico’s Massachusetts Playbook. This gives us a great, natural experiment in how quickly elected officials follow accounts with newly increased relevance to their work. We see that Lisa is followed by 26% of the State House, but 42% of City Council.

While not a perfect proxy, when combined with data on who follows media outlets like WBZ and the Boston Globe where the difference is much less, this indicates that City Councillors use Twitter to keep up with news more than state legislators do and are more likely to hit the follow button when a reporter takes over an importance publication.

There are three State House Twitter communities and two Boston communities

Network analysis gives us the ability to detect communities within messy social networks. 

To understand this concept, imagine we had a list of which Members of Congress followed which professional sports teams on Twitter. Even if we hid the names of the members, we would see clusters emerge: a group that follows @RedSox, @Patriots, @Celtics, and @NHLBruins; another that follows @Yankees, @Mets, and so on. This would tell us there’s a New England community, a Tri-State community, and more throughout the country. 

The sports team example could be analyzed manually with enough time and effort. But what if we want to see what communities emerge from hundreds of accounts, or tens of thousands of connections? 

Luckily, there are algorithms we can apply to a large network (in this case, 26,960 accounts followed by Boston elected officials), and let it make its best guess at sorting communities. 

Applied to the Massachusetts State House, this process detects three communities. One is mostly Democrats in the Progressive Caucus, the second is mostly Democrats who are not in the Progressive Caucus and also includes a majority of the Republicans, and the third is almost entirely comprised of Republicans. 

This suggests that Beacon Hill has a large Progressive group, a large moderate group, and a smaller Republican group, with overlap and movement between. To the extent that this reflects political reality is, of course, a matter for debate - this is only one data point.

When we apply the same community process for Boston-specific politics, we find two: roughly one city and one state.The first community is almost entirely members of the State House with Boston districts, and the second contains 10 of the 12 councillors.

Interestingly though, about a third of the State Representatives do fall into the “City” community, whereas only a fifth of the State Senators do. 

Does this mean that Reps are more likely to be plugged into city-based informational networks, or that their smaller district sizes mean they follow more of the local voices that councillors follow? The smaller geography may mean that Reps’ day-to-day priorities are more tied to the Councillors than fellow Members.

Boston-based media is followed in Boston

We would of course expect that legislators follow media based in their districts. Nevertheless, it was surprising to see how many more Boston-based voices were followed by elected officials from the city than by those across the state. 

Some of the more interesting points:

• More City Councillors follow the Dorchester Reporter than the Boston Globe.

• A number of the most highly-followed reporters and columnists are 30 percentage points more followed in Boston than by legislators across the Commonwealth. Some of the highest, which reflect both a Boston focus and high Boston attention, include Bill Forry (+49 percentage points), Universal Hub (+51 percentage points), and Chris Lovett (+55 percentage points).

• BNN News is followed by two-thirds of the State House delegation with districts inclusive of Boston, but by almost no state legislators from outside the city, showing its local, but extensive reach. 

At-large candidates are unevenly followed by electeds

If the percentage of Boston electeds following a candidate translated to votes, next year’s At-Large Council seats would include Michael Flaherty, Julia Mejia, Ruthzee Louijeune, and David Halbert.

Of course it’s not that simple, and this is not a prediction of who will win. But it will be interesting to check after the election to see how Twitter matches up with what the voters decide. It may be that there’s no connection and that these follows don’t reflect the larger electorate. 

It could also be that elected officials following a candidate on Twitter is an early indicator of established political connections, support, and name recognition. This could be used to parse out a race before polling and endorsements are available.

Do mayoral messages come across online?

Here is the same information for the mayoral candidates, with former Mayor Marty Walsh included for comparison.

Of the mayoral candidates, Michelle Wu has by far the most total Twitter followers, whereas the highest percentage of Boston politicians follow Annissa Essaibi George. However, all candidates have high percentages of Boston followship, perhaps because Twitter is an information battleground in a high profile race. These leading candidates tweet multiple times per day, sending out hundreds of messages per month.

Looking into what they tweet, we can use a comparison cloud to see which words are disproportionately used by each candidate. 

Annissa Essaibi-George is the most likely to tweet “getbosbacktobiz,” “citykidswin” and “students”; Kim Janey is the leading user of “history,” “herstory,” “love,” and “proud”; John Barros tweets “affordable” the most. Andrea Campbell is the candidate who most uses “equitable,” “accountability,” and “justice,” while Michelle Wu is the most likely to say “community,” “bold,” “change” and “climate”. 

We don’t yet know whether the candidates have been able to get their messages across online, or whether any has definitively carved out ownership on particular issues, but this gives us some insight into their relative messaging priorities. 

Here’s a comparison of their Twitter feeds:

How Legislata helps

Legislata is productivity software for people in politics. Streamline your emails, better manage your tasks, and save time that can be spent crafting your message and policies. We’ll be launching in September. Get in touch to register your interest or subscribe for more information and analysis like this. Follow us on Twitter at @LegislataApp for more insights.