Growth of user generated content and online social media has led to interesting problems in computational modeling of user behavior. One such problem is to predict reactions to social content posted online. This becomes increasing important due to the surge in the usage of social media by business enterprises and political parties to either create brand awareness or form opinions. The prediction model is learned from past reactions of the community to similar posts and used for prediction on a new post as well as providing recommendations to get better reactions to the social post.