Hadoop Application Architectures

Chapter 8. Clickstream Analysis

Clickstream analysis commonly refers to analyzing the events (click data) that are collected as users browse a website. Such an analysis is typically done to extract insights into visitor behavior on your website in order to inform subsequent data-driven decisions.

You can use clickstream analysis to figure out how users interact with your online presence, which geographic regions or time zones your traffic originates from, what devices and operating systems are most commonly used to access your site, what the common paths are that your users take before they do something that’s of interest to you (usually referred to as a goal—e.g., register for the mailing list, sign up for an account, or add an item to the shopping cart), and so on. You can correlate such click data with your marketing spend to do further analysis like the return on investment (ROI) of various marketing channels (organic search, paid search, social media spending, etc.). You can also, optionally, join click data with your ODS (Operational Data Store) or CRM (Customer Relationship Management) data to do further analysis based on additional information about your users.

Although this chapter uses analysis on clickstream data to show how to put together the various pieces described in Part I for a given use case, the concepts described here are applicable to any “batch” processing system that operates on machine-generated data. Examples of such data include but are not limited to ...