pith. sign in

arxiv: 2509.11444 · v1 · submitted 2025-09-14 · 💻 cs.CL · cs.SI

CognitiveSky: Scalable Sentiment and Narrative Analysis for Decentralized Social Media

Gaurab Chhetri , Anandi Dutta , Subasish Das This is my paper

Pith reviewed 2026-05-18 16:05 UTC · model grok-4.3

classification 💻 cs.CL cs.SI
keywords Blueskysentiment analysisdecentralized social mediatransformer modelsnarrative analysisemotion detectioncomputational social scienceAPI data ingestion
0
0 comments X

The pith

CognitiveSky applies transformer models to Bluesky posts to enable scalable sentiment, emotion, and narrative analysis on decentralized platforms.

A machine-rendered reading of the paper's core claim, the machinery that carries it, and where it could break.

The paper introduces CognitiveSky as an open-source framework that pulls data from Bluesky's API and applies transformer-based models to label user content for sentiment, emotion, and narratives. These labels produce structured outputs that feed a dynamic dashboard tracking shifts in emotions, activity levels, and conversation topics. The system runs entirely on free-tier infrastructure to maintain low costs and broad access. It is demonstrated on mental health discourse but built with a modular design to support other uses such as disinformation detection and crisis monitoring. The work positions this approach as a bridge between large language models and emerging decentralized networks for computational social science.

Core claim

CognitiveSky ingests Bluesky data through the platform API, applies off-the-shelf transformer models to annotate posts for sentiment, emotion, and narrative categories, and generates structured summaries that drive real-time visualization of evolving discourse patterns in a dashboard. The framework achieves low operational cost by relying on free infrastructure and is shown operating on mental health topics while remaining extensible to other domains through its modular structure.

What carries the argument

The CognitiveSky framework, which combines Bluesky API data ingestion with transformer model annotations to produce structured outputs and dashboard visualizations.

If this is right

  • Researchers gain a tool for real-time monitoring of public discourse on federated platforms like Bluesky.
  • The modular design allows the same pipeline to support applications in disinformation detection and civic sentiment tracking.
  • Low-cost operation on free infrastructure broadens access for computational social science studies.
  • Structured outputs enable further quantitative analysis of trends in emotion and narrative across user-generated content.

Where Pith is reading between the lines

These are editorial extensions of the paper, not claims the author makes directly.

  • The same ingestion and annotation approach could be adapted to compare discourse patterns across multiple decentralized networks.
  • If the annotations prove reliable, the dashboard could serve as an early indicator for shifts in public mental health discussions during events.
  • Adding platform-specific fine-tuning steps might address potential mismatches between general models and Bluesky-style short posts.

Load-bearing premise

Off-the-shelf transformer models can deliver sufficiently accurate and unbiased annotations for sentiment, emotion, and narratives on Bluesky posts without domain adaptation or reported validation.

What would settle it

A direct comparison of the framework's automatic annotations against independent human labels on a sample of Bluesky posts, measuring agreement rates for sentiment and emotion categories.

Figures

Figures reproduced from arXiv: 2509.11444 by Anandi Dutta, Gaurab Chhetri, Subasish Das.

Figure 1
Figure 1. Figure 1: End-to-end overview of CognitiveSky’s modular pipeline. [PITH_FULL_IMAGE:figures/full_fig_p003_1.png] view at source ↗
Figure 2
Figure 2. Figure 2: Interactive visual components from the CognitiveSky dashboard. Panels include (a–b) word clouds for [PITH_FULL_IMAGE:figures/full_fig_p004_2.png] view at source ↗
read the original abstract

The emergence of decentralized social media platforms presents new opportunities and challenges for real-time analysis of public discourse. This study introduces CognitiveSky, an open-source and scalable framework designed for sentiment, emotion, and narrative analysis on Bluesky, a federated Twitter or X.com alternative. By ingesting data through Bluesky's Application Programming Interface (API), CognitiveSky applies transformer-based models to annotate large-scale user-generated content and produces structured and analyzable outputs. These summaries drive a dynamic dashboard that visualizes evolving patterns in emotion, activity, and conversation topics. Built entirely on free-tier infrastructure, CognitiveSky achieves both low operational cost and high accessibility. While demonstrated here for monitoring mental health discourse, its modular design enables applications across domains such as disinformation detection, crisis response, and civic sentiment analysis. By bridging large language models with decentralized networks, CognitiveSky offers a transparent, extensible tool for computational social science in an era of shifting digital ecosystems.

Editorial analysis

A structured set of objections, weighed in public.

Desk editor's note, referee report, simulated authors' rebuttal, and a circularity audit. Tearing a paper down is the easy half of reading it; the pith above is the substance, this is the friction.

Referee Report

2 major / 1 minor

Summary. The manuscript introduces CognitiveSky, an open-source and scalable framework for sentiment, emotion, and narrative analysis on the decentralized social media platform Bluesky. It ingests posts via the Bluesky API, applies transformer-based models to produce structured annotations, and drives a dynamic dashboard for visualizing evolving patterns in emotion, activity, and topics. The system is built on free-tier infrastructure for low cost and accessibility, demonstrated for mental health discourse monitoring, and positioned as modular for applications including disinformation detection and crisis response.

Significance. If the annotations prove reliable, CognitiveSky could provide a transparent, extensible tool for computational social science on federated platforms, bridging LLMs with decentralized networks. Strengths include its open-source design, emphasis on accessibility and modularity, and focus on real-time discourse analysis in shifting digital ecosystems. However, the absence of any quantitative validation, benchmarks, or error analysis substantially limits its assessed significance as a deployable research instrument.

major comments (2)
  1. [Abstract / System Description] Abstract and system description: the central claim that CognitiveSky offers a reliable tool for computational social science rests on the accuracy of direct application of off-the-shelf transformer models to Bluesky posts, yet no domain adaptation, held-out validation sets, accuracy metrics, or error analysis are reported. This is load-bearing because platform-specific features (abbreviations, threading, discourse styles) may cause systematic mislabeling that propagates to all downstream visualizations and analyses.
  2. [Abstract] Abstract: the assertions of scalability and low operational cost on free-tier infrastructure lack supporting benchmarks for throughput, latency, or resource usage when processing large-scale Bluesky data. Without these, the performance claims for real-time analysis remain unverified.
minor comments (1)
  1. The manuscript would benefit from explicit section headings and a system architecture diagram to clarify the data flow from API ingestion through model annotation to dashboard output.

Simulated Author's Rebuttal

2 responses · 0 unresolved

We thank the referee for their constructive and detailed comments on the manuscript. We address each major comment below and have revised the paper to incorporate the feedback where it strengthens the work.

read point-by-point responses

  1. Referee: [Abstract / System Description] Abstract and system description: the central claim that CognitiveSky offers a reliable tool for computational social science rests on the accuracy of direct application of off-the-shelf transformer models to Bluesky posts, yet no domain adaptation, held-out validation sets, accuracy metrics, or error analysis are reported. This is load-bearing because platform-specific features (abbreviations, threading, discourse styles) may cause systematic mislabeling that propagates to all downstream visualizations and analyses.

    Authors: We agree that annotation reliability is important for downstream use in computational social science. The manuscript presents CognitiveSky as an open-source framework and demonstration system rather than a newly validated model; it applies existing transformer models to Bluesky data to enable real-time visualization. We have revised the abstract and added an explicit limitations section that discusses the risks of domain shift, platform-specific language, and the absence of new held-out validation or error analysis in this work. The section recommends user caution and outlines plans for future domain adaptation. This change directly addresses the concern while preserving the contribution of the modular, accessible pipeline. revision: yes

  2. Referee: [Abstract] Abstract: the assertions of scalability and low operational cost on free-tier infrastructure lack supporting benchmarks for throughput, latency, or resource usage when processing large-scale Bluesky data. Without these, the performance claims for real-time analysis remain unverified.

    Authors: We acknowledge that quantitative performance data would better support the scalability claims. In the revised manuscript we have added a performance subsection reporting observed throughput (posts per minute), end-to-end latency for ingestion and inference, and resource consumption on the free-tier services during the mental-health monitoring deployment. These metrics confirm feasibility for real-time operation at the scales shown, with discussion of modular scaling options for larger volumes. The addition provides the requested evidence without changing the core accessibility focus. revision: yes

Circularity Check

0 steps flagged

No significant circularity in framework description

full rationale

The paper describes the construction and application of an open-source software system (CognitiveSky) that ingests Bluesky posts via API and applies off-the-shelf transformer models for sentiment, emotion, and narrative annotation. No mathematical derivations, equations, fitted parameters, or predictions are claimed. The central contribution is an engineering integration and dashboard, with no load-bearing steps that reduce by construction to the paper's own inputs, self-citations, or ansatzes. This is a self-contained systems paper whose results do not rely on any internal circular reduction.

Axiom & Free-Parameter Ledger

0 free parameters · 1 axioms · 0 invented entities

The contribution rests on standard NLP assumptions about model transfer rather than new theoretical entities or fitted parameters.

axioms (1)
  • domain assumption Pre-trained transformer models can be applied directly to annotate sentiment, emotion, and narratives in Bluesky social media text.
    This assumption underpins the core annotation step described in the abstract.

pith-pipeline@v0.9.0 · 5696 in / 1198 out tokens · 45651 ms · 2026-05-18T16:05:17.392130+00:00 · methodology

discussion (0)

Sign in with ORCID, Apple, or X to comment. Anyone can read and Pith papers without signing in.

Lean theorems connected to this paper

Citations machine-checked in the Pith Canon. Every link opens the source theorem in the public Lean library.

What do these tags mean?
matches
The paper's claim is directly supported by a theorem in the formal canon.
supports
The theorem supports part of the paper's argument, but the paper may add assumptions or extra steps.
extends
The paper goes beyond the formal theorem; the theorem is a base layer rather than the whole result.
uses
The paper appears to rely on the theorem as machinery.
contradicts
The paper's claim conflicts with a theorem or certificate in the canon.
unclear
Pith found a possible connection, but the passage is too broad, indirect, or ambiguous to say the theorem truly supports the claim.

Forward citations

Cited by 1 Pith paper

Reviewed papers in the Pith corpus that reference this work. Sorted by Pith novelty score.

  1. Retrieval-Augmented Generation Must Move Beyond Factual Grounding to Represent Diverse Opinions

    cs.AI 2026-04 unverdicted novelty 5.0

    Opinion-aware RAG with LLM opinion extraction and entity-linked graphs improves retrieval diversity by 26-42% over factual baselines on e-commerce forum data.

Reference graph

Works this paper leans on

3 extracted references · 3 canonical work pages · cited by 1 Pith paper

  1. [1]

    Al Hariri, Y ., Chausson, S., Ross, B., & Magdy, W. (2024). Twixplorer: An interactive tool for narrative detection and analysis in historic twitter data.Companion Publication of the 2024 Conference on Computer-Supported Cooperative Work and Social Computing, 83–86. Barbieri, F., Camacho-Collados, J., Espinosa Anke, L., & Neves, L. (2020). TweetEval: Unif...

    work page 2024

  2. [2]

    (2025).Oracle cloud free tier

    Corporation, O. (2025).Oracle cloud free tier. https:// www.oracle.com/cloud/free/ Das, S., & Dutta, A. (2020). Characterizing public emotions and sentiments in covid-19 environment: A case study of india [Published online: 14 July 2020].Journal of Human Behavior in the Social Environment. https : //doi.org/10.1080/10911359.2020.1781015 Das, S., Dutta, A....

    work page doi:10.1080/10911359.2020.1781015 2025

  3. [3]

    W., & Iamnitchi, A

    Ng, K. W., & Iamnitchi, A. (2023). Coordinated information campaigns on social media: A multifaceted framework for detection and analysis.Multidisciplinary International Symposium on Disinformation in Open Online Media, 103–118. Pew Research Center. (2025, April). Teens, social media and mental health. https://www.pewresearch. org/internet/2025/04/22/teen...

    work page 2023