Analysis of Bigmart Outlets sales using Big Data Analytics Method

Sales prediction is an important problem for businesses. A sales forecast can guide a company's production, pricing, and promotion strategies, and can help mitigate risks associated with under or overproduction. Traditional methods of sales prediction utilized statistical techniques like regression analysis or time series analysis, but these methods have certain limitations, including the need for large and clean data sets. This research work focuses on exploring the patterns and trends present in a large-scale retail dataset through exploratory data analysis (EDA). The primary objective of this research is to uncover insights that can enhance the understanding of customer behaviour, identify opportunities to improve business processes and inform data-driven decisions. The data consists of daily sales records for 1560 products of ten BigMart Outlets. This study explores several preprocessing steps, such as handling missing values and outliers, scaling and normalization, and feature engineering. The findings also revealed that the top-selling items were bought by a diverse range of customer demographics, indicating that marketing campaigns should target a broader audience. Furthermore, the study identified several unusual transactions that required further investigation.