Topic Modelling (and more) with NLP framework Gensim
summary
Demo shows a complete pipeline: spaCy preprocessing (lemmatization, curated stopwords, bigrams) turns messy news text into a gensim-ready corpus; gensim runs LSI / LDA / HDP models with a bag‑of‑words dictionary; pyLDAvis gives interactive topic overlap and the tutorial computes coherence scores (example ~0.56 vs 0.50) to compare models. The actual claim is that cleaning changes results more than fiddling with the model and the provided Jupyter notebooks reproduce the process end to end.
AI-written summary · about
chapters (9)
- 0:00Topic modeling with Gensim
- 2:11Text preprocessing importance
- 6:50Spacy vs NLTK for cleaning
- 9:37Clean corpus with Spacy
- 19:52Bigrams for topic coherence
- 23:29Text to bag-of-words
- 26:24First Gensim topic models
- 36:00Visualize topics with pyLDAvis
- 41:55Key takeaways and next steps
related talks
01h 29mTricks, tips and topics in Text Analysis
01h 01mWorkshop: Intro to topic modeling using Python
01h 38mNavigating the World of NLP and Text Analytics
01h 31mModern NLP in Python
01h 19mGenerating Customers Insights with Topic Modelling and HuggingFace SetFit Method
01h 26mBagging to BERT- A tour of applied NLP
25mInventing Curriculum using Python and spaCy
35mHow does text become data?