Mobile cell data-based mobility forecast: applying neural networks to enrich human mobility analysis We created mobility networks using a national-scale telecommunications provider's comprehensive network data, which comprised 365 days of telecommunication traffic data in 2019. The predictive power of the presented forecast method, based on mobile cell data, is similar to those based on navigation data with a significantly finer spatial resolution. The advantage of applying a mobile cell data-based forecast model is that it enriches the predictive model with descriptive data since it also embraces user and device data. The mobility graphs were trained with neural network to predict the presence values and expected mobility in 30-minute time slots.