Retail Product Data Analysis

Project Overview: This project focuses on exploratory data analysis (EDA) and business insights generation using a retail product dataset containing missing values. It focuses on identifying, analyzing, and handling missing data in a retail product dataset. Various imputation techniques were applied based on data distribution and business reasoning rather than blanket filling. It demonstrates end-to-end data analysis skills using Python, SQL, and Power BI.

Dataset: Retail Product Dataset with Missing Values

• Records:4361 entries distributed in 5 columns category, price, rating, discount, stock
• The dataset intentionally includes missing values to simulate real-world data challenges.
• Source: Kaggle - https://www.kaggle.com/datasets/himelsarder/retail-product-dataset-with-missing-values

Data Cleaning & Preprocessing: The following steps were performed using Python (Pandas):

• Identified missing values in numerical and categorical columns

• Imputed: Numerical columns using median values Categorical columns using mode values Removed duplicate records Verified data types and corrected inconsistencies Created derived fields where necessary for analysis

• Exploratory Data Analysis (EDA) Numerical Analysis: Distribution analysis using histograms with KDE; Outlier detection using boxplots. Categorical Analysis: Category-wise product count; Stock availability comparison (In Stock vs Out of Stock); Category-level price comparison using boxplots; Correlation Analysis - Pearson correlation between: Price & Discount, Price & Rating, Discount & Rating.

• Result: All correlations are close to zero, indicating no strong linear relationship between pricing, discounts, and ratings.

SQL Analysis SQL was used to answer business-oriented questions such as: Total number of products

• Average price, discount, and rating
• Products with high discounts (≥ 40%)
• Category-wise pricing trends
• Stock availability breakdown
• The SQL queries are consolidated and included for easy review and reproducibility.

Power BI Dashboard An interactive Power BI dashboard was created to visualize:

• Price and discount distributions (histograms)
• Category-wise price comparison
• Stock availability
• Correlation summary using DAX measures

Key Highlights:

• Custom DAX measures for correlation
• Clean layout with business-focused KPIs
• Designed for stakeholder-friendly interpretation

Tools & Technologies used:

• Python: Pandas, NumPy, Matplotlib, Seaborn
• SQL: SQLite
• Power BI: DAX, interactive dashboards
• Version Control: Git & GitHub

Key Insights

• Most products fall within a mid-price range with moderate discounts
• Category C dominates the product count
• High discounts do not necessarily correspond to higher ratings
• Price, discount, and rating operate largely independently in this dataset

Conclusion This project showcases practical data analysis skills, including data cleaning, EDA, SQL querying, and dashboard creation. It reflects real-world scenarios where data is imperfect and insights must be derived through structured analysis.