Approaches And Tools For Visualizing Data Journalism(NBA Data)

Nowadays, we have entered the era of massive data production, dissemination and applica-tion. Public available data appears increasingly on the Internet, making the traditional news media industry turned into "Data Journalism". Data journalism denotes the news reports driven by da-ta, evidence and truth, and seeks to depict the background, evolution, causality of news based on interactive visual representation, which is no doubt more persuasive and exploratory than the tra-ditional way of news generation. However, with the increasing amount of data, the higher data dimensions, and the more complicated data sources, how to effectively process and visualize the data has become a challenging problem.This thesis is tightly centered with visualization for data journalism. Aiming at the field of sports news, this thesis proposes an effective news visualization approach for basketball games for the purposes of quick review of the game procedure and insightful analysis of game play. The visualization system presents an overview of all games that allows for single game selecting and shows a single game visualization that illustrates game procedure, scores and play actions for further explorations. The implemented system provides a means of understanding basketball games, and allows for interactive analysis of behavior of individual player. The experiments show that data journalism is more convincing and exploitable than traditional text-based news report.Additionally, data journalism is field-oriented, with data varies in types and structures. To make the traditional news practitioners easily implement news visualization, we developed an open source visualization JavaScript library, called DataV.js, cooperated with visualization team of Taobao.com company. It provides 8 visual forms: treemap, node-link diagram, stack graph, scat-terplotmatrix, force-directed graph, bubble charts, matrix and chord diagrams. DataV.js provides an easy-to-use interface and configuration options, facilitating non-professional users generate vi-sualizations conveniently. DataV.js has been widely used in many applications.