Description

Online advertising have risen in the last few years and with that comes the need of reaching the accurate audience and the necessity of serving the right advertisement to the right customer. In the world of online advertising, decisions must be made with low latency. Therefore, ad click prediction is a specific problem and is an example of high-velocity and high-throughput problem.

As Computer science and Management students and potential data scientists, our aim in this project was to build an ad click prediction system which processes an inputted data set of real world user data and produces a file enhanced by a first column stating if the user is likely to click on a displayed ad or not -represented by a Boolean value (true or false).

Technologies

The tools and programming languages used for this project are :

  • Dataiku
  • Scala
  • Spark