Aleksandar Tomašević🛠️ transforEmotion crossed 25K CRAN downloads! Paper is out in Computational Communication Research.
Open-source R package for emotion analysis across text, images, and video. Everything runs locally on CPU, no APIs, no cloud.
New: local RAG with small LLMs, VAD scoring, vision model registry, built-in benchmarking, uv for zero-config Python environments.
https://www.aup-online.com/content/journals/10.5117/CCR2026.2.2.TOMA
https://github.com/atomashevic/transforEmotion
#RStats #OpenScience #OpenSource #EmotionAnalysis #NLP #FOSS
transforEmotion: An Open-Source R Package for Emotion Analysis Using Transformer-Based Generative AI Models | Amsterdam University Press Journals Online
This software demonstration article introduces transforEmotion, an open-source R package that addresses critical bottlenecks in communication research emotion analysis. Communication researchers currently face three key barriers: (1) modal fragmentation requiring separate tools for text/image/video analysis, (2) rigid emotion taxonomies that don't match theoretical frameworks, and (3) irreproducible workflows dependent on commercial APIs. transforEmotion removes these bottlenecks through unified multimodal processing, zero-shot classification with arbitrary labels, local inference capabilities, and seamless R integration. This enables systematic investigation of emotion dynamics across communication contexts and modalities that were previously technically difficult to analyze.
Nicola Mustone