Geo-contextual TV Consumption Patterns using Unsupervised Learning methods

Author(s): Harshil Agrawal, Nitin Vinayak Agrawal, Shubham Gupta, Mrigank Shekhar

Abstract

According to the Adobe Digital Insights Advertising Report [3], 74% of Americans feel that the television commercials they view are irrelevant to them. This is a huge proportion of the total audience, which also shows the number of capital advertisers is wasting on advertising to uninterested audiences. Creating TV personas will help target relevant audiences, thus optimising the ad campaigns. In present times, with an ever-increasing television demand, advertisers are more and more interested in studying television content from the consumer’s perspective. TV communities emerging due to this consumption are crucial for digital marketing planners since it gives a clearer view of who the brands’ audience could be, what they like, and what they do. This would enable the planners not only to understand the avenues to reach them but also to understand what creative messaging works best, what social messaging works best and eventually better returns. This paper examines and defines television consumers’ behaviour to gain a more in-depth understanding of the target audience. As the Digital world is progressing towards a cookie-less future, i.e., cookie-level information won’t be available, we ought to look into aggregated data at the geo level. The unsupervised clustering experiments (using KMeans) are performed on TV data provided by major TV players in the market. In this paper, we shall walk you through the initial steps of data preparation and techniques used for clustering. In the latter half, we will discuss more the techniques, experiments and the interpretation of TV personas with data.

The Chartered Data Scientist Designation

Achieve the highest distinction in the data science profession.

Explore more from Association of Data Scientists

Become ADaSci Chapter Lead

As a chapter lead, you will have the opportunity to connect with fellow data professionals in your area, share knowledge and resources, and work together to advance the field of data science.