Date of Award

5-2024

Document Type

Project

Degree Name

Master of Science in Information Systems and Technology

Department

Information and Decision Sciences

First Reader/Committee Chair

Dr. Conrad Shayo

Abstract

This culmination project investigated and analyzed the impact of factors influencing sales forecasting models in the e-commerce ecosystem. The research questions are: Q1) To what extent are sales forecasting models in the e-commerce ecosystems influenced by customer demographic variables such as gender, age, and geographic locations? Q2) To what extent are sales forecasting models in the e-commerce ecosystems influenced by product specific factors? The datasets used were from Kaggle, and the Worldometer websites. The findings are: Q1) Individuals aged '55 or over' significantly impact total sales in both the USA and Brazil. Male consumers consistently accounted for a higher proportion of total sales compared to female consumers and were primary contributors across all age brackets in both countries. The concentration of top-performing states and customers is in regions of China and the USA. Q2) The top 10 product categories, including "Outwear and Coats", "Jeans", and "Sweaters", significantly contribute to revenue generation, accounting for 68.21% of total sales. The top 20 recognizable brands such as "Diesel", "Calvin Klein", and "The North Face" collectively contribute 22.04% of total sales. The XGBoost model outperformed the linear regression model with an R-square value of 0.98. The conclusions are: Q1) Customer demographic factors significantly impact e-commerce sales forecasting models. Analyzing sales across age, gender, and geography unveils crucial consumer behavior trends. This underscores the necessity of targeting specific demographic segments and tailoring marketing strategies accordingly. Q2) Customer demographics significantly impact e-commerce sales forecasting. Analyzing sales across age, gender, and location reveals valuable consumer behavior patterns, necessitating targeted marketing and inventory strategies. Further studies include exploration of: (a) customer behavior dynamics in response to segmentation variables, and cultural influences on preferences. (b) Investigation of cyclical trends in sales forecasting models, and (c) utilizing advanced machine learning techniques, like deep learning algorithms to improve predictive accuracy.

Download

  Contact Author

Share

COinS