The number of phishing websites on the world wide web is steadily increasing as a popular attack vector towards potential victims. Phishing websites can be differentiated by legit websites using a select set of identifiers like content, URL and the use of SSL certificates. In this research, two websites with visual similarity are built: one resembles a legit website, the other one contains specific characteristics of phishing website (and thus resembles a phishing version of the earlier mentioned website). With a custom built API, mouse movements, mouse clicks and key presses were captured on both websites to record user behaviour. With this captured data, which resulted in more than 100.000 records, data science models were applied to find statistical relevance. And the results? Attend This Session And You Will Be Surprised!