E-commerce sales forecasting using web-scale data

Abstract

Peekd’s dataset encompasses the largest Google Analytics panel worldwide, includingsales data with customer information from 150k unique websites. Although vast, peekd does not cover the entire internet. Peekd is looking to expand its data analytics to include sales predictions to enhance its offering to clients. A subset is used containing only European consumer electronics websites. The goal is to predict the websites transactions for the next seven days. This serves as a proof of concept for peekd to roll out the sales prediction feature as quickly as possible.