A Predictive Visual Analytics Method for Taxi Routines

This paper presents a predictive visual analytics method for exploring taxi trajectory data.First,a new algorithm that integrates traditional LSTM with time information and weather conditions is designed to predict the traffic in the city.Second,a trip planning model based on real-time forecasting data is proposed,in which the constraints of generalized location semantics are included for more reasonable result.Finally,a suite of visualization tools is developed to analyze the taxi operation and predict their income in each road,which can enhance the operational decisions of drivers.Experimental results demonstrate the effectiveness of our method.