🧠 Introduction

This app applied natural language processing (NLP) to large sets of qualitative data, helping researchers and psychologists identify themes and emotional tone within interviews.

🎯 Objectives

• Cluster semantically similar responses using transformer embeddings
• Provide interactive visualization for exploring response clusters
• Integrate sentiment and topic summarization
• Enable comparison of language models within a single interface

⚙️ Methods

• Implemented using R Shiny and reticulate to run sentence-transformers
• UMAP and clustering (K-means / HDBSCAN) for dimensionality reduction
• Interactive visualization via plotly and ggplot2
• Model benchmarking for multilingual data

💡 Outcome

Delivered a powerful NLP visualization platform enabling qualitative researchers to move from manual coding to data-driven insights. The app demonstrated how machine learning can assist psychological interpretation.