The application of machine learning in spam detection has been implemented since 1990s for emails. The shift to social networks over email for communication has allowed for data to be viewed in many dimensions, aside from the core content. Reactions to content, such as likes, and metadata about the creator, allows us to gain a larger context for user created content. This project aims to study the impact in effectiveness of spam filtering, with the available context of social networks, specifically Facebook. In addition to studying the effectiveness of the result, the project will be a learning experience in machine learning, and feature engineering. The wider context of the target data also increases the challenge of feature engineering. Strategies ranging from basic, naive filtering, to strategies used in production, such as online filtering are covered.