Abstract:
"The COVID-19 pandemic has changed how people shop online, making it important for
businesses to find new ways to predict customer buying patterns. This project aims to analyse
and predict customer behaviour in e-commerce after the pandemic using data from 2019
onward. This project uses advanced data analytics, including machine learning models, to
uncover useful insights from past purchase data. This study focuses on forecasting sales
quantities over time to understand customer purchase behaviour better. By looking at sales
quantities, it is evident that trends and patterns can be seen in what and how much customers
are buying, which helps businesses plan better. Tools such as Python, Orange, and Weka were
used, and time series analysis was applied with algorithms such as Long Short Term Memory
(LSTM), Autoregressive Integrated Moving Average (ARIMA), Holt-Winters, Exponential
Smoothing (ETS), Simple Moving Average (SMA) and Weighted Moving Average (WMA).
The goal is to provide recommendations for e-commerce businesses to improve customer
satisfaction, marketing, and operations. By understanding how sales quantities change over
time, businesses can make better decisions and adapt to the evolving digital markets. This
research shows how predictive analytics can help businesses grow during these challenging
times."
