Start Date
16-8-2018 12:00 AM
Description
This paper describes our analysis and findings from a “big data†project associated with an international travel and tourism company obtained through the Wharton Customer Analytics Initiative (WCAI). Our dataset contains a large number of consumer clickstreams, their historical purchases, and the advertisements they were exposed to during their online activity among other attributes. Our analysis based on Markov Chain Model finds that a clickstream with a long click history does not necessarily provide better prediction than a clickstream with a memory of just one click. This finding is important because it has a potential to simplify the clickstream-based prediction models as well as lessen the concern for consumer privacy because of the non-significance of cookies in predicting the next click.
Recommended Citation
Bhandari, Gokul; Bussiere, David; and Voyer, Peter, "How Loud is the Scream of a Clickstream? Insights from Big Data Analysis" (2018). AMCIS 2018 Proceedings. 41.
https://aisel.aisnet.org/amcis2018/DataScience/Presentations/41
How Loud is the Scream of a Clickstream? Insights from Big Data Analysis
This paper describes our analysis and findings from a “big data†project associated with an international travel and tourism company obtained through the Wharton Customer Analytics Initiative (WCAI). Our dataset contains a large number of consumer clickstreams, their historical purchases, and the advertisements they were exposed to during their online activity among other attributes. Our analysis based on Markov Chain Model finds that a clickstream with a long click history does not necessarily provide better prediction than a clickstream with a memory of just one click. This finding is important because it has a potential to simplify the clickstream-based prediction models as well as lessen the concern for consumer privacy because of the non-significance of cookies in predicting the next click.