A Big Data Analysis System for Use in Vehicular Outdoor Advertising

Abstract

Outdoor advertising is an old industry and the only reliably growing advertising sector other than online advertising. However, for it to sustain this growth, media providers must supply a comparable means of tracking an advertisement’s effectiveness to online advertising. The problem is a continual and emerging area of research for large outdoor advertising corporations, and as a result of this, smaller companies looking to join the market miss out on providing clients with valuable metrics due to a lack of resources. In this paper, we discuss the processes undertaken to develop software to be used as a means of better understanding the potential effectiveness of a fleet of private car, taxi or bus advertisements. Each of the steps present unique challenges including big data visualisation, performance data aggregation and the inherent inconsistencies and unreliabilities produced by tracking fleets using GPS. We cover how we increased the metric aggregation algorithm performance by roughly 20x, built an algorithm and process to render data heat maps on the server side and how we built an algorithm to clean unwanted GPS ‘jitter’.