CASE STUDY

Click Prediction

anCompany Overview

Location Intelligence technology company.

Collecting data from ~ 80M active unique devices per month.

Business Case

Predict whether an ad will be clicked.

 

Data

Big data – ~1.5M rows each day, 62 features.

Data is comprised of

  • Lower Funnel: spend, impressions, clicks, installs, etc.
  • Targeted Audience: country, age range, IP address.
  • Ad data: name of app and app category.

Tools

C++ machine learning platforms for speed and scalability.

Business Impact and Insights

Machine learning model outperformed previous models developed by the company.

Feature importance analysis provided valuable business insights regarding location intelligence.

< BACK TO projects portfolio