QVC is an American cable, satellite and broadcast television network with a flagship shopping channel specializing in televised home shopping. The problem domain was to analyze QVC’s data to identify top selling products and product categories, effect of email campaigns and provide a model to forecast sales.
Dataset has 6 tables (details in data dictionary), with 4.7+ million records after joining relevant tables.
Questions:
- Q1 - Identify top selling products and product categories?
- Q2 - Are email campaigns effective in driving sales? How much effective?
- Q3 - Create a model to forecast Sales
Check Executive Summary for Analysis Results.
Check Jupyter Notebooks for Analysis Steps.
Skillset Used - Python (pandas, matplotlib, scikit-learn), R, Data Wrangling, Feature Engineering, Time Series, ARIMA, ARIMAX, Random Forest, Gradient Boosting
#TBD - Apply LSTM to Q3