Inferring users’ geographical locations is an essential task for various internet-based applications such as advertising, recommendation systems, social event detection, and emergency localization. Despite the massive volume of content being exchanged online, particularly on social media platforms, only a limited portion includes references to the users’ geographical locations. This emphasizes the necessity of utilizing available data to predict user geolocation. Therefore, we propose a framework composed of a Neural Network and a Random Forest capable of determining one geographic location based on a collection of social media interactions. Experiments conducted on a real-world benchmark dataset demonstrate that this model outperforms baseline methods.