Walmart-sales-Analysis-Dashboard

This project analyzes Walmart sales data to uncover key business insights and forecast future sales. It combines data cleaning, exploratory analysis, feature engineering, and forecasting techniques with deployment of a lightweight web app.

The work demonstrates the end-to-end data science process:

• Cleaning raw transactional data
• Performing Exploratory Data Analysis (EDA)
• Creating a cleaned dataset for modeling
• Building forecasting models for weekly sales trends
• Deploying an app interface for interaction with the analysis

---

Business Objectives

1. Identify trends and patterns in Walmart’s historical sales data.
2. Analyze the impact of holidays, promotions, and seasonal events on sales.
3. Forecast weekly sales using predictive modeling.
4. Build a simple web app for showcasing predictions.
5. Provide insights for inventory and staffing decisions.

---

Project Structure

├── app.py # Flask/Streamlit app for deployment ├── ArjunP_24202600_WalmartSales.ipynb # Jupyter notebook with analysis ├── Uncleaned_Walmart_Sales_Data.csv # Raw Walmart dataset ├── Cleaned_Walmart_Sales_Data.csv # Preprocessed dataset

---

Data Description

The Walmart dataset includes:

• Store – Store ID
• Date – Weekly sales date
• Weekly_Sales – Sales amount (target variable)
• Holiday_Flag – Whether the week includes a holiday
• Temperature – Average weekly temperature
• Fuel_Price – Fuel cost in the region
• CPI – Consumer Price Index
• Unemployment – Regional unemployment rate

---

Methodology

1. Data Cleaning

   • Removed nulls, handled duplicates
   • Standardized column values
   • Created Cleaned_Walmart_Sales_Data.csv

2. Exploratory Data Analysis (EDA)

   • Trends across time, stores, and holidays
   • Correlation between macroeconomic variables (Fuel Price, CPI, Unemployment) and sales

3. Modeling & Forecasting

   • Time-series forecasting for weekly sales
   • Comparisons of different forecasting methods

4. App Development

   • Built a web app (app.py) for simple predictions and visualization
   • Uses Flask/Streamlit for deployment

---

Key Insights

• Holiday weeks show significantly higher sales spikes.
• Unemployment and CPI negatively correlate with sales.
• Store-level analysis reveals variance in sales performance across locations.
• Forecasting models highlight seasonality and long-term trends.

---