Emotions are a fundamental part of humans and influence their decisions, actions and interactions. Accurately identifying emotions has a wide range of applications from healthcare to human computer interaction. The paper provides a comprehensive exploration of EEG-based emotion recognition, elucidating its basics, surveying existing studies, and highlighting the central role of machine learning (ML) in achieving accurate classification. The machine learning analysis employs k-Nearest Neighbours (kNN), Support Vector Classifier (SVC), and Multi-layer perceptron (MLP) models. Refinement strategies involving the categorization of electrodes into distinct regions and hyperparameter tuning using the PyCaret library lead to a substantial accuracy improvement, with Random Forest Classifier achieving 67% accuracy for valence and Ridge Classifier achieving 64% accuracy for arousal. These dimensions of valence and arousal serve as key indicators, mapping to up to two distinct emotions.