Rossmann Retail Sales Analysis on Snowflake

This repository contains the end-to-end SQL-based data analysis of the Rossmann store sales dataset, performed entirely within the Snowflake cloud data platform. The primary goal of this project was to conduct data cleaning, feature engineering, and exploratory data analysis (EDA) to uncover insights that could inform a demand forecasting model.

Key Objectives

• Data Processing at Scale: To manage and process a large dataset with over 1 million rows efficiently using Snowflake's cloud architecture.
• Feature Engineering: To create new, meaningful features from existing data to improve the quality of the analysis.
• Insight Discovery: To perform EDA to identify key drivers of sales and understand the impact of factors like promotions, holidays, and seasonality.

🛠️ Tech Stack

• Cloud Data Platform: Snowflake
• Language: SQL
• Dataset: Rossmann Store Sales (Kaggle)
• Visualisation: Tableau Public

Methodology

The entire project followed a structured analytical process within Snowflake:

1. Data Cleaning & Preparation

• Loaded the train and store datasets into Snowflake tables from a stage.
• Conducted a thorough analysis to identify and handle NULL values through logical imputation (e.g., filling missing CompetitionDistance with the median).
• Normalized categorical columns, such as converting StateHoliday from 0 to 'None' for clarity.

2. Feature Engineering

• Joined the train and store tables to create a comprehensive primary analysis table.
• Engineered new time-based features by extracting the Year, Month, Week, and Day from the date column.
• Created custom boolean flags like IsWeekend and PromoMonthFlag to better capture seasonal and promotional effects.

3. Exploratory Data Analysis (EDA)

• Wrote a series of SQL queries to analyze the relationships between various features and sales.
• Investigated the distribution of store types, assortment levels, and the impact of competitor proximity.
• Analyzed the effect of promotions, holidays, and day-of-the-week on average sales.

📊 Key Insights

One of the most significant findings from the EDA was the quantifiable impact of promotions on sales:

• The average daily sales for a store on a non-promotional day were €4,406.
• During promotional periods, the average daily sales surged to €7,991.
• This represents a quantifiable sales uplift of 81.5%, highlighting the critical success of the promotion strategy.

📊 Interactive Dashboard

The SQL analysis from Snowflake has been visualized in an interactive Tableau Public dashboard, covering the four key insights from the EDA:

• Promo vs Non-Promo Sales — 81.5% uplift (€4,406 → €7,991 avg daily sales)
• Sales by Store Type — performance comparison across store categories A/B/C/D
• Monthly Sales Trend — seasonality pattern across the full year
• Holiday Impact — effect of public, Easter & Christmas holidays on sales

🔗 View Live Dashboard on Tableau Public

📈 Visualizations

Below are some screenshots of the key queries and the resulting analysis performed in Snowflake.

1. Sales Data

2. Sales Bucket VS Frequency Graph

3. Month wise Avg. Sales Data

4. School Holiday Sales Data

5. State Holiday Sales Data