Uploaded August 2025 | Updated September 2026, 2 weeks ago
📺 Predicting Game of Thrones Deaths Using Network Science 🧠⚔️
What if you could predict which Game of Thrones characters would die — before the final seasons aired?
In this talk, we dive deep into the fascinating world of network science meets fantasy TV. Starting with subtitles from the HBO series (not the books!), we reconstruct a complete character co-occurrence network — mapping who appears with whom, how often, and how closely connected they are across all six seasons.
🔍 Then, using simple machine learning, we turn this network into a dataset of features — like node degree and centrality — and train a binary classifier to predict who’s most likely to die next.
💡 Spoiler alert: the results were eerily accurate. One listener even used the model to win bets in their MBA cohort!
🔗 Whether you're into data science, storytelling, or just love Game of Thrones, this is a powerful example of how structured thinking and messy data can reveal hidden patterns — even in Westeros.
📺 Predicting Game of Thrones Deaths Using Network Science 🧠⚔️
What if you could predict which Game of Thrones characters would die — before the final seasons aired?
In this talk, we dive deep into the fascinating world of network science meets fantasy TV. Starting with subtitles from the HBO series (not the books!), we reconstruct a complete character co-occurrence network — mapping who appears with whom, how often, and how closely connected they are across all six seasons.
🔍 Then, using simple machine learning, we turn this network into a dataset of features — like node degree and centrality — and train a binary classifier to predict who’s most likely to die next.
💡 Spoiler alert: the results were eerily accurate. One listener even used the model to win bets in their MBA cohort!
🔗 Whether you're into data science, storytelling, or just love Game of Thrones, this is a powerful example of how structured thinking and messy data can reveal hidden patterns — even in Westeros.
