Metro Crowd Prediction Using Deep Learning and User Location

Metro systems, being public transportation systems, have been under strain due to growing urbanization. Effective management of crowd in metro systems is crucial to avoid congestion and enhance the overall experience of commuting. This paper presents a deep learning-based method for predicting crowd density in the metro system based on historical ridership and user location information. We use a convolutional neural network in tandem with location-based data to predict the crowd levels at the different stations highly accurately for better service planning and passenger flow management.