DECODING E-COMMERCE FLUCTUATIONS: A MACHINE LEARNING ANALYSIS OF INFLUENTIAL VARIABLES DURING US COVID-19 (2010-2024)

This research examines factors influencing the US e-commerce market during the Covid-19 crisis, investigating consumer behavior across three periods: pre-pandemic (2010–December 2019), pandemic (December 2019–2021), and post-pandemic crisis (2022–2024). Using multiple linear regression analysis in Python and Machine Learning techniques, the study evaluates the impact of key economic indicators (Gross Domestic Product, Unemployment Rate, Consumer Price Index, Internet Penetration Rate, and Consumer Sentiment Index) on e-commerce sales. These variables were used to develop a mathematical model explaining the relationship between economic and sentiment indicators and e-commerce growth. During the pandemic, e-commerce activity surged as lockdowns forced consumers to rely more on online shopping. Post-pandemic, as restrictions eased and confidence recovered, the market exhibited continued growth, surpassing pre-pandemic levels. Despite the initial surge driven by restrictions, e-commerce remained strong even after their removal. The analysis highlights the importance of economic factors in shaping e-commerce trends, with GDP and CPI emerging as particularly influential. Additionally, the study underscores the critical role of internet penetration in sustaining e-commerce, especially during physical distancing measures. These findings provide insights into how economic and technological factors drive long-term changes in consumer behavior within the e-commerce sector.