The rising number of electric vehicles (EVs) is expected to cause congestion at charging stations and strain the charging capacity, potentially impacting power grids. This work attempts to address this difficulty by forecasting the charging schedules for residential charging stations using machine learning (ML) and deep learning (DL) models. Historical weather and traffic data, frequently disregarded in conventional EV load forecasts, were used to make the projection realistic. This work attempted to forecast session duration and energy consumption using data from a large housing cooperative in Norway. For energy consumption, the mean absolute error (MAE) of 0.329400 and root mean squared error (RMSE) of 0.489502 were recorded for random forest as the best result. Similarly, gradient boosting offered the best results for session duration with an MAE of 0.276741 and RMSE of 0.435624. On the other hand, the transformer method yielded an MAE of 0.483480 and an RMSE of 0.781634 for energy consumption, whereas for session duration, the metrics were 0.467384 and 0.744851. Therefore, based on these results, ML models provided superior results over the transformer model.