• 中国计算机学会会刊
  • 中国科技核心期刊
  • 中文核心期刊

计算机工程与科学

• 论文 • 上一篇    下一篇

基于Movielens电影数据的可视分析

许冰晗,尚鸿运,马灿,李尚   

  1. (天津大学软件学院,天津 300354)
  • 收稿日期:2016-01-04 修回日期:2016-06-27 出版日期:2017-11-25 发布日期:2017-11-25

Visual analysis of Movielens data

XU Bing-han,SHANG Hong-yun,MA Can,LI Shang   

  1. (School of Computer Software,Tianjin University,Tianjin 300354,China)
  • Received:2016-01-04 Revised:2016-06-27 Online:2017-11-25 Published:2017-11-25

摘要:

目前,随着电影数据逐渐被人们获取,关于电影数据的研究可以给人们带来很多启发。分析电影流派的演变规律,可以为导演提供电影题材建议;分析经济和电影之间的关系,可以找到电影演变的原因;研究高评分电影在时间上的规律,可以指导导演选择电影的上映时间。但是,由于电影包含电影名称、所属流派、评分等多重属性,一般的研究方法不足以发现并直观地呈现电影数据隐含的规律。用可视化与可视分析的方法分析电影数据,设计了一系列相互关联的可视化视图,从多个时间尺度角度分析电影流派的时间演变,通过增长率曲线图研究电影数量和经济的相关关系,并设计饼图集来发现高评分电影在时间、流派上的规律。
 

关键词: 电影, 经济, 可视化与可视分析

Abstract:

As we can collect increasingly growing movie data, study on movie data brings some inspiration. For example, analysis of the evolution of film genres can provide directors with recommendations on movie themes, and analysis of the relationship between economy and movie can find some reasons of film evolution. Study on high-score movies’ regularity over time can guide directors to choose movies’ release date. However, general research methods are inadequate to discover and present patterns hidden behind movie data visually because a film contains a movie name, genre information, ratings and other multiple attributes. We therefore adopt a visualization and visual analysis method for analyzing films and design a series of associated visualization views, which can analyze the time evolution of films in each genre from multiple temporal scales. We can also discover the correlation between movies and economy through the growth curve, and design a pie view to find patterns of high-score movies over time and across genre.

Key words: movie, economy, visualization and visual analysis