Abstract: To create traffic trajectory visualization through visual graphics interactively,the traffic trajectory-oriented data flow visualization method is proposed.The method uses wizard recommendations to assist in data flow construction for traffic track clustering,correlation analysis and anomaly detection.Firstly,the data flow system is developed to interactively create traffic trajectory visualization.Secondly,the wizard recommendation model based on heterogeneous networks of data flow is proposed to assist in data flow construction.The recommendation model treats wizard recommendations as a graph search problem and feeds user ratings on wizard recommendations to a heterogeneous network of data flow to update node relationships.The results of the case evaluation based on real traffic trajectory data and the evaluation of the Likert scale on users’propensity to choose indicate the validity of the approach.