Development of an intelligent agent for real-time recommendation of sport analytics in live broadcasted events

该项目是设计和开发一个智能数据平台，以推荐相关的背景信息图形，以覆盖在转播的体育赛事的视频馈送上。目标是自动化数据处理和图形生成任务，并设计一个基于先前生成的序列进行训练的推荐系统。因此，推荐者将在观看体育比赛时根据什么被认为是有趣的有价值的信息来提供相对的序列。系统的调整和最终的编辑决定留给操作员。该项目的目标是：(1)收集、讨论、分析和指定此类平台所需的数据源；(2)原型、验证，然后开发智能推荐系统，以实现数据处理和图形分类模块的自动化；以及(3)使系统能够适应新的数据源，并根据平台计划实现的主要目标整合数据变化。该项目的成果是智能推荐系统的原型。该项目是与AstuceMedia合作完成的，AstuceMedia是一家总部位于蒙特利尔的公司，专门从事实时互动广播图形设计和数据集成。

The project is the design and development of an intelligent data platform to recommend relevant contextualinformational graphics to overlay on the video feed of broadcasted sport events. The goal is to automate thedata processing and graphic generation task and design a recommender system that is trained on the previousgenerated sequences. The recommender will thus provide relative sequences according to what is consideredinteresting valuable information while viewing sports games. The adjustments of the system and the finaleditorial decision is left to the operator. The objectives of the project are: (1) to collect, wrangle, analyze, andspecify required data sources needed for such a platform; (2) prototype, verify, and then develop an intelligentrecommender system to automate the data processing and graph classification modules; and (3) empower thesystem to adapt with new sources of data and integrate data changes according to the main goals the platform isplanned to achieve. The deliverable of the project is a prototype of the smart recommender system. This projectis done in partnership with AstuceMedia, a Montreal-based company specialized in real-time interactivebroadcast graphics design and data integration.