📘 Introduction

1 What Drives Online Purchases?

The visualization below shows how the number of product-related pages viewed and the value of pages visited affect the likelihood of completing a purchase.

2 🛍️ Summary:

This project explores how online shopping behavior predicts whether someone completes a purchase.

We used data from the UCI Machine Learning Repository, collected by the Faculty of Computer Science and Engineering at Gazi University (2018). The goal? To uncover which behaviors — like product page views, bounce rates, and session duration — actually signal a shopper is likely to buy.

🧪 We modeled purchase behavior using a logistic regression, where the outcome was whether a person made a purchase (yes or no).

Key predictors in our model:

🔎 Product Pages Viewed

⏳ Time Spent on Site

🚪 Bounce Rate (leaving the site quickly)

💡 Insight: People who view more product pages are about 35% more likely to make a purchase — even after accounting for time spent on the site and whether they bounced (with a margin of error of ±3%).