@Marinelivesorg — Project Biography

Quarterly topic evolution across 10,142 original tweets · July 2012 – June 2025
NMF topic model (15 topics) with TF-IDF bigrams

Topic Composition Over Time

Each band shows a topic's share of quarterly tweet content. Click legend items to show/hide topics.

Individual Topic Trajectories

Select a topic to see its trajectory and example tweets. Hover for quarter-by-quarter values.

Topic Index

Click any topic to highlight its trajectory above and see representative tweets.

Quarter × Topic Heatmap

Cell intensity shows each topic's share within that quarter. Read across rows to see a topic's life cycle; read down columns to see what dominated each quarter.
Method: Topic extraction via Non-negative Matrix Factorisation (NMF) on TF-IDF vectors with unigrams and bigrams. 10,142 original tweets (retweets excluded), filtered to ≥5 words after preprocessing (URLs and @mentions removed). 15 topics, min document frequency 5, max document frequency 50%. Quarterly aggregation by mean topic weight across all tweets in each quarter.