Amazon Sales Data Analysis

Project Overview

This project involves a comprehensive Exploratory Data Analysis (EDA) of Amazon sales data. The goal is to derive actionable business insights regarding order trends, product popularity, fulfillment methods, and customer geography. The analysis focuses on data cleaning, processing, and summarizing key metrics using Python.

Dataset Description

The dataset contains 128,975 records and 23 columns. It includes transactional data such as:

• Order Details: Order ID, Date, Status, Amount.
• Product Details: SKU, Category, Size, Style.
• Fulfillment: Ship Service Level, Fulfilled By, Courier Status.
• Customer Location: Ship City, Ship State, Ship Country.

Tech Stack

• Language: Python
• Libraries: Pandas
• Environment: Jupyter Notebook

Data Cleaning & Preparation

Before analysis, the raw data underwent significant preprocessing:

1. Handling Null Values:
   • Dropped the empty Unnamed: 22 column.
   • Filled missing Amount values with 0.
   • Filled missing Courier Status with "Unshipped".
   • Dropped rows with missing city/state locations.
2. Data Type Conversion:
   • Converted the Date column to datetime format to extract monthly trends.

Key Insights

1. Sales Performance

• Top Month: April was the highest-performing month with ~28.8 Million in revenue, followed by May (~26.2M) and June (~23.4M).
• March Anomaly: March shows significantly lower data points (171 orders), likely indicating partial data availability for that month.

2. Product Analysis

• Top Categories: "Set" and "Kurta" are the dominant categories, accounting for the vast majority of orders.
• Revenue Drivers: The "Set" category generates the highest revenue (~39M), followed by "Kurta" (~21M) and "Western Dress" (~11M).
• Size Preferences: M, L, and XL are the most popular sizes among customers.

3. Fulfillment & Logistics

• Fulfillment Authority: Amazon fulfills significantly more orders (approx. 70%) compared to Merchant fulfillment.
• Service Level: The majority of orders use Expedited shipping (~88k orders) over Standard shipping (~40k orders).

4. Geography

• Top States: The highest number of orders originate from Maharashtra and Karnataka, followed by Telangana and Uttar Pradesh.

5. Customer Type

• B2B vs B2C: The platform is overwhelmingly B2C (Business to Consumer), with B2B transactions making up less than 1% of the total orders.

Future Improvements

• Visualization: Integrate Matplotlib or Seaborn to visualize sales trends and geographic heatmaps.
• Code Optimization: Refactor the monthly analysis sections into reusable functions to reduce code repetition.
• Statistical Modeling: Implement forecasting to predict sales for future months.

How to Run

1. Clone the repository.
2. Install the required libraries:

   pip install pandas

3. Load the Amazon Sale Report.csv file into the project directory.
4. Run the main.ipynb (or Amazon_Sales_EDA.ipynb) file.