Online shopping behavior is the process by which consumers search for, select, purchase, use, and dispose of goods and services, over the internet. For the ecommerce platform, one of the most important questions is, whether the customer is just browsing or actually buying.
Customers are very heterogenous so it is important that the sellers do not treat them in the same way. They always want to leverage their resources to find and keep the customers in which they have confidence that they can more likely to purchase. The sellers could take some proactive action, like time-limited coupons or free trials, to push customers to purchase. By targeting the right customers the sellers could improve retention and increase sales and profits.
In this project we did an intensive analysis of consumer behavioral and performed the following tasks:
- Classification
- Exploratory Data Analysis (EDA)
- Modelling
- Explainability with SHAP
- Customer Segmentation
- KMeans Clustering
- Semi-Supervised Learning
- Label Spreading
The source of the data in the project is Online Shoppers Purchasing Intention Dataset.
The complete code can be found in this notebook and this Python script.
- Tools/techniques used: Python, Jupyter Notebook, Scikit-learn, Seaborn, Matplotlib, Pandas, Numpy, Scipy, SHAP
- Algorithms used: Logistic Regression, Random Forest, KMeans, Label Spreading
GitHub Repository for this project.