Following Cavafy in social media (part II)

BigVisualData
5 min readMay 24, 2024

--

In the first part of this series we dealt with the theoretical background behind my research. Let us now consider the case of Cavafy and his appropriation by popular culture, namely in the ways social media users appropriate the poet and his work.

Anna-Maria Sichani, (2015), in one of the first attempts to map Cavafy on the Web, noted:

“The methods by which social web users interpret, approach and engage with Cavafy enrich and reinvigorate the readers’ role, by bringing to the forefront concepts such as user-generated content, interaction, networking and sharing mechanisms.”

Photo posted on Instagram, a user holding a copy of Cavafy’s poems, pondering over what she has just read

The Poet as Pop Icon

The corpus of this part consists of Twitter data collected on April 30th, 2023, one day after the date of his birth and death. Search was conducted querying for

Καβάφης OR Καβαφης OR Cavafy OR Kavafis

We collected 283,166 tweets and retweets, since January 1st, 2010 (4,867 days), an average of 58 tweets and retweets per day.

Tweets on Cavafy per day for 4.8K days. See the interactive version here.

When contemplating this diagram one should always keep in mind that it is also depicting the slow growth of Twitter in Greece as well as globally. Therefore, we are not that much interested in numbers than in patterns. And there is a clear pattern in this diagram:

There is a constant interest in Cavafy, but not too high. There is an average of 58 tweets and retweets per day since January 1st, 2010. The actual number is lower at the beginning of the period and higher after 2017 or so. But the pattern is mostly related to the spikes. Can you guess what are they related with?

April 29th, the date of Cavafy’s birth and death. [A sort of Obituary]

March 21st, the World Poetry Day. [An opportunity to show off cultural capital]

Other spikes are of lesser importance, and they are not recurrent. In any case, it is possible to try and understand the context.

For example, the very first spike is related to the decision by the leader of the far-right LAOS party to leave the governmental coalition before voting for the second memorandum in early February 2012. When he left the meeting he recited some verses from “Che fece . . . . il gran rifiuto”.

The most intriguing tweet was, in my opinion:
“Αν ήξερε ο Καβαφης τι τον περίμενε ξημέρωμα 9ης Φεβρουαρίου θα είχε γίνειΚαρυωτακης”, i.e., “If Cavafy knew what would happen on February 9th, he would have become Karyotakis”.

Another spike is found on October 31, 2020, the day Sean Connery died. Seven hundred users worldwide tweeted with a link to a Youtube video where the late Sean Connery is reciting Cavafy’s Ithaca, over music by Vangelis.

The Network

When people interact in social media, when they reply, quote, retweet, or mention another user, then they become part of a network, extending across borders, beyond the personal circle of friends and acquaintances. When two people (nodes) interact with the same user (node), they also are connected through this third user. And when more and more users interact with each other often, they create a communicative community within the wider network, a sub-network or sub-graph. Users who often interact with others, who are energetic within the network, they form a lot of connections, of edges or lines leading from them to the others. Others may be less active, but act as the focus of communication, they are significant because all others turn to them, mention them, or quote and retweet their content. The number of incoming and outgoing edges to a node is called its “degree.” In-degree designates the number of incoming edges, while out-degree is the number of outgoing edges.

Other characteristics are also important in a network. Nodes who are low in both in- and out-degree may still get importance because they are close to the most significant nodes of the network. Others act like hubs, connecting communities that would otherwise be independent, thus becoming the bridges keeping the network together.

There are a lot to learn from such a network. Collecting the interactions between users around the name of the poet, we found that more than 100,000 users tweeted about Cavafy. More than 98,000 are represented in the network (some may have erased their tweets, or their accounts, or made them private).

The Cavafy network on Twitter (X).

The users who tweeted most were:

CCavafy, an English speaking account posting small pieces from Cavafy’s diaries of the same date (day-month) and old photos, usually of places mentioned in the post. Aaron Timms, who presented at Cavafy Summer School in 2022, mentions at length his meeting with the @CCavafy account.

DurrellSociety, well, is the International Lawrence Durrell Society.

Paul Holdengraber, is Former Founding Dir. Onassis LA & Dir. LIVEfromNYPL | Interviewer, Instigator, Curator of Public Curiosity | Also Quotomania & TheQuarantineTapes

The rest top tweeting users were usually posting poetry and literature.

It is possible to realize that the communicative communities found in the network, and indicated with different colors are mostly also linguistic communities.

The major communicative communities are linguistic ones.

The collected tweets contain approximately 1.6 million words. The 870 most repeated words are forming a co-occurence network, meaning a network where edges indicate that two words (nodes) exist in close proximity within the same tweet.

The co-occurence network of the top 870 words within the tweets. A tripartite structure is clear, with each constituent composed by more communities

The linguistic communities found to the users’ network are found here as well. They are aligned behind the different transcriptions of the poet’s name. Italian, Spanish and Turkish related to Kavafis, English to Cavafy, and Greek to Καβάφης.

--

--

BigVisualData

Analyzing Visual Corpora with computational methods. It’ll provide pieces on methodology, sociological & semiotics viewpoints. Yannis Skarpelos