Superstore Sales Analysis Project

Introduction

The Superstore Sales dataset provides a comprehensive view of sales data over a period, covering various product categories, geographic regions, and customer segments. This project focuses on extracting insights and building predictive models to better understand sales trends and make informed business decisions.

Dataset Overview

The dataset includes sales data from a superstore, featuring the following key attributes:

• Order Date: Date when the order was placed.
• Ship Date: Date when the order was shipped.
• Product Category: Category of the product sold.
• Sales Region: Geographic region where the sale occurred.
• Customer Segment: Type of customer who made the purchase.
• Shipping Mode: Mode of shipping used for delivery.
• Sales Amount: Total sales value of the transaction.

The dataset contains both numeric and categorical features, offering a rich source of information for detailed analysis.

Goal of Analysis

The primary goal of this project is to gain actionable insights into sales trends, patterns, and drivers within the superstore. By exploring the data, we aim to:

• Identify key factors influencing sales performance, such as product categories, geographic regions, customer segments, and shipping modes.
• Develop predictive models to accurately forecast future sales.

Importance of Insights

Understanding sales dynamics is crucial for optimizing inventory management, resource allocation, and marketing strategies. By uncovering insights from the data, the superstore can:

• Identify high-performing product categories.
• Target lucrative customer segments.
• Optimize shipping operations.
• Anticipate demand fluctuations.

These insights enable data-driven decision-making, leading to improved operational efficiency, customer satisfaction, and overall business performance.

Model Evaluation

After evaluating the performance of our predictive models, the ARIMA model stands out as the top performer for predicting Superstore product sales.

• Mean Square Error (RMSE): ~ 243407.191
• Mean Absolute Error (MAE): 431.358
• R-squared: 0.605 (indicating that around 60% of the variance in sales data is captured by the model).

These results highlight that ARIMA suits better the dataset due to the timeline in the sales.