This Web App is a comprehensive data analysis dashboard designed for analyzing Supermarket sales. It offers detailed views for Customer Relationship Management (CRM), Sales Trends Analysis, and Product Analysis. Additionally, a fine-tuned LLM-based AI chat feature is in the works and will be available soon.
View the live application at: Supermarket Analytics Dashboard
This aims to help sales and management answer questions such as who are our quality customers, what are they buying and how can we get them to buy more and build a relationship with our customers?
Read the code walkthrough here on Medium
This project is developed using:
Python for scripting and data manipulation.
MySQL and
PostgreSQL for relational database management.
- Pandas for data manipulation and analysis.
- Seaborn and Matplotlib for data visualization.
- Scikit-learn for implementing machine learning models.
- Joblib for model serialization and deployment.
- Streamlit for the backend and UI of the application.
- Supabase for handling database interactions.
- Streamlit Community Cloud for hosting the application.
- Daily Customer Insights: View and analyze daily customer interactions, purchase behavior, and overall engagement.
- Customer Segmentation: Easily segment customers based on their purchasing patterns and other criteria.
- Sales Performance: Track and visualize sales performance over time, identifying peaks, trends, and seasonal behaviors.
- Revenue Breakdown: Get a clear view of revenue contributions from different product categories.
- Top Products: Identify and analyze the best-performing products in the supermarket.
- Product Segmentation: View sales performance and other metrics for specific product segments.
- This app features a content-based filtering product recommendation system, leveraging data-driven insights to suggest products that customers are likely to purchase based on their past behaviors and preferences. The recommendation engine is built using:
- Pandas for data processing.
- Scikit-learn for implementing the content-based filtering algorithm.
- Joblib for model persistence and fast retrieval.
- View the notebook used to train this model here
- A fine-tuned language model (LLM) will be integrated into the app, providing an AI-powered chat feature. This AI assistant will specialize in sales data analysis and provide insightful responses related to customer trends, product performance, and overall supermarket analytics.
Data is based on real supermarket data courtesy of a supermarket in Kenya, personal information has been randomized for preview.
... View more at Supermarket Analytics Dashboard
Icons used in this project are courtesy of geticon.