Table 5 The pipeline summary Step Method... | Emerald Publishing

Table 5

The pipeline summary

Step	Method	Purpose	Output
Preprocessing	Tokenization, lemmatization (spaCy)	Clean and normalize text	Standardized corpus
Topic modelling	LDA (Gensim)	Identify latent themes	Topic distributions
Clustering	TF-IDF + PCA + k-means	Group similar discourse segments	Semantic clusters
Key-Phrase extraction	RAKE	Extract salient ESG terms	Keyword sets
Sentiment analysis	VADER	Measure evaluative tone	Sentiment scores
Digital framing	Keyword detection	Identify digital mediation	Binary/cluster-level tagging
Validation	Coherence, silhouette, expert review	Ensure robustness	Validated outputs

Step	Method	Purpose	Output
Preprocessing	Tokenization, lemmatization (spaCy)	Clean and normalize text	Standardized corpus
Topic modelling	LDA (Gensim)	Identify latent themes	Topic distributions
Clustering	TF-IDF + PCA + k-means	Group similar discourse segments	Semantic clusters
Key-Phrase extraction	RAKE	Extract salient ESG terms	Keyword sets
Sentiment analysis	VADER	Measure evaluative tone	Sentiment scores
Digital framing	Keyword detection	Identify digital mediation	Binary/cluster-level tagging
Validation	Coherence, silhouette, expert review	Ensure robustness	Validated outputs