The goal of this project is to analyze the sales, customer behavior, inventory management, supplier performance, and profitability of WideWorldImporters to derive actionable insights. The project involves data exploration, analysis, and modeling to address several business problems and optimize different aspects of operations.
֎ Live Dashboard:
֎ Tools & Technologies used:
» SQL : for querying the WideWorldImporters database.
» Python : for data analysis and visualization (using Pandas, NumPy, Matplotlib, Seaborn).
» Machine Learning : for sales forecasting, customer segmentation, and predictive analytics.
» Data Visualization : Power BI for dashboarding and reporting.
֎ Methodology:
-
1. ARIMA - Sales Forecasting
2. KMeans Clusting - Customer Segmentation
3. BetaGeoFitter - Customer Lifetime Value
4. APRIORI - Market Basket Analysis (Association Rules)
» Data Extraction (SQL)
» Data Cleaning & Transformation (PowerQuery)
» Data Exploration and Preprocessing (Python)
» Data Analysis (Excel)
» Machine Learning and Predictive Modeling (Python)
» Data Cleaning (PowerQuery)
» Visualization and Reporting (PowerBi)
֎ Data Modeling:
֎ The following questions were addressed in the project (Project Scope):
» Sales Performance Analysis : Which product categories have the highest sales performance?
» Sales Forecasting : Can we predict future sales for the next quarter based on historical data?
» Customer Segmentation : How can we segment customers based on their purchasing behavior (e.g., frequency of purchase, average purchase value, or product preferences)?
» Inventory Optimization : Are there any products that are consistently overstocked or understocked, and how can we better manage inventory?
» Customer Lifetime Value Prediction : Can we predict the future lifetime value of a customer, and how can we identify high-value customers?
» Market Basket Analysis (Association Rules) : What are the common product associations in customer orders (i.e., what products are frequently purchased together)?
» Profitability by Product/Category : Which product categories or individual products contribute the most to profit margins?
֎ Findings & Results:
-
1. Segment 0: Customer with higher purchase frequency and lower purchase value.
2. Segment 1: Customer with higher purchase frequency and higher purchase value.
3. Segment 2: Customer with lower purchase frequency.
» Packaging Materials and Clothing are the categories with the highest Sales Performance
--------------------------------------------------------
» Future sales predicted for the next quarter based on historical data
--------------------------------------------------------
» Customer was assigned to 3 different segments:
--------------------------------------------------------
» Overstocked Product: Shipping carton (Brown) 305x305x305mm.
» Understocked Product: 3 kg Courier post bag (White) 300x190x95mm.
--------------------------------------------------------
» Future lifetime value of a customer
--------------------------------------------------------
» The purchase association between products is very low for the majority of the products. Below is a list of the products with highest probability to be purchased together.
--------------------------------------------------------
» Highest selling Product: Air cushion machine (Blue).
» Highest selling Category (after unifying the Novelties subcategories and the Clothing subcategories into the parent categories): Novelties.
