Posts with No Response: The Island of Loneliness

Loneliness and isolation are eternal emotions in human beings. Technological advancements create ample avenues, like social media, for individuals to articulate themselves and record emotions. Yet the sense of loneliness has never vanished, as these expressions are easily buried in the digital stream.

We analyze tweets that express loneliness during holiday seasons but receive few responses. By superimposing digital charts on physical models, we visualize these lonely posts and generate the island of loneliness. The work aims to reveal the complexities of human emotions in the digital age and to reflect on the interconnections between technology, solitude, and social communication.

Social media platforms generate a constant flow of expressions, opinions, and emotions. Much attention goes to opinion leaders and posts that receive extensive responses; few works explore the unacknowledged expressions in this landscape. Like grains of sand sinking to the bottom of the data sea, unanswered posts form an unseen layer that merits attention.

We focus on posts that convey loneliness. The absence of responses amplifies the sense of isolation and compels us to shed light on these silent voices. They record human loneliness and indicate its perpetual presence [1, 3, 4, 7]: even in a world where the Internet connects people and everything, individuals can still experience deep loneliness.

From roughly a 1% sample of the Twitter stream between December 23 and January 3 of 2016 to 2019, we collected 162,376,361 tweets. Filtering for posts containing “lonely” or “loneliness” outside proper nouns, and keeping only those with no response, left 22,133 lonely tweets.

A Latent Dirichlet Allocation model [2, 6] with 20 topics and an emotion-detection model (Emotion English DistilRoBERTa-base) [5] turn each tweet into a vector of weighted topics and emotions. This vector drives the visualization.

Social media is an expansive sea of semantic data into which people continuously throw their thoughts and emotions [8]. Each post is a stone cast into the sea. A few carry enough weight to generate continuous ripples and spark mass responses. The majority are lightweight and unremarkable; like grains of sand they sink to the seabed, occasionally resurfacing when a keyword is searched, but mostly existing as unnoticed sediment.

Loneliness contour. Isomap reduces each tweet’s vector to two dimensions; proximity in the resulting scatter plot means thematic similarity. The density of tweets, drawn as contours, delineates the shape of the island, and adjusting the bandwidth simulates the island’s accumulation. In the projection, a sequential scale from dark to bright blue encodes height.

Emotion side view. An emotion degree combines positive and negative emotion values. The upper part of the side view shows the island’s height; the lower part shows the emotion degree, where darker grey and lower position mean a more negative emotion. Like the submerged part of an iceberg, the negative emotions behind lonely tweets are easily overlooked.

Noisy ripple. Ripples capture the bustling surface of the water. Their radius encodes the response degree, the sum of reposts, likes, and comments. The noisy ripples are deliberately juxtaposed with the quiet sedimentation of unanswered posts.

The loneliness contour and emotion side view are rendered with D3.js. To deepen the solid presence of loneliness, the contour at bandwidth 8 is physicalized: it is turned into a three-dimensional model in Rhino, refined in AutoCAD, laser-cut from wood, and assembled by hand. Ripples and contours are projected over the island model through time. Touching the model extends the encounter beyond the visual channel.

Social media is a unique intersection between public information and private life. Not every post is shared with the expectation of a response; sometimes people simply use the platform as an outlet. The absence of engagement does not mean these expressions are insignificant. It highlights the complex dynamics of social media and the diverse motivations behind sharing personal experience. Whether people who feel lonely seek responses when they express it remains an open question.