Back to Projects

Natural Language Processing

NLP Text Analysis and Topic Modeling

A coursework-style NLP notebook project combining GloVe embeddings, word similarity analysis, representation bias discussion, spaCy preprocessing, and LDA topic modeling. I completed this notebook work myself as formal ML/NLP practice.

Period

May 2026

Tools / Tech

PythonJupyter NotebookScikit-learnspaCyGensimGloVeLDA

Why I built it

I wanted to connect classic NLP ideas with hands-on notebooks: vector representations, preprocessing choices, topic-word distributions, and document-level interpretation.

Links

What it includes

  • Explores GloVe embeddings and word similarity.
  • Uses spaCy preprocessing before topic modeling.
  • Builds and interprets LDA topic models through word-topic and document-topic views.

What I worked on

  • Prepared notebook workflows for embeddings, similarity checks, preprocessing, and topic modeling myself.
  • Discussed representation bias and limitations of embedding-based analysis.
  • Interpreted topic model outputs rather than only reporting model artifacts.

What I Learned

  • Learned how preprocessing decisions affect downstream NLP interpretation.
  • Practiced connecting mathematical representations to readable explanations.
  • Built a more cautious view of embedding similarity and topic labels.