To help users to determine the most appropriate visualization type is a useful feature of business visualization tools. Existing systems often give preliminary suggestions based on data sources but usually cannot make practical final decision. User stories are generalizations of user requirements. To recommend visualization type based on user stories can make better use of human experience to achieve automated decision making. One approach discussed in the paper is using machine learning techniques to model existing visualization types with corresponding user stories, and then use this model to predict recommended visualization type for new user story. This paper designs and implements a recommendation system prototype ReViz to verify the feasibility of this approach. As a typical web application, Modeling, Input Processing and Predicting components of ReViz are programmed using Python with Flask framework and Anaconda package set, and user interface is implemented using HTML, JavaScript and CSS with Bootstrap front-end library. The evaluation results show that ReViz can give recommended visualization type based on user story keywords. As a data-based intelligent software development technology achievement, visualization type recommendation system can also be integrated into larger business information management systems.