Prediction of Propensity For Click Through Rate – Big Data Solution

Project Description

The business problem is about predicting the propensity score of clicks on various advertisements for a leading advertisement corporation in digital market industry.

Key challenge is the processing of click through rate data is difficult to perform on standalone system due to large volume of data. Also it leads to Increase in cost of infrastructure due to space usage and processing time.

The solution proposed include:

  • Hadoop is used to distribute the big data across various systems connected
  • Cloudera is used for managing the big data with integration of R and Hadoop
  • R is used to as an analytical tool to build the business model based on CRISP-DM
  • The customer was greatly benefitted in terms of:

  • Reduced costs as all the tools are open source
  • Better efficiency due to increased processing of big data in real time