Graph ML / Data Engineering
Twitch Language Detection
Predicted language communities from graph structure using embeddings + supervised learning.
Stack
PythonNode2VecXGBoostNetworkX
Contributions
- ·Generated Node2Vec embeddings
- ·Trained and tuned XGBoost classifier
- ·Built evaluation and visualization pipeline
Impact
90.1%accuracy on 168k-node Twitch SNAP graph
0.88macro F1 across six language families