Sales Trend and Forecasting Using ML Final Year Project

• Secure admin login
• Admin JWT token stored in cookie
• Admin dashboard
• View customer statistics
• View revenue statistics
• View subscription plans
• Quick navigation links
• Manage subscription plans
• Create subscription plan
• Edit subscription plan
• Delete subscription plan
• Add plan name
• Add plan price in INR
• Manage ML actions
• Create ML feature/action
• Edit ML action
• Delete ML action
• Map each ML action to a subscription plan tier
• Manage packages
• View user package assignments
• Edit user package assignments
• Manage user upgrades
• Manage customers
• List registered users
• Search customers
• View customer details
• Manage registered users
• View payment and revenue records
• Track subscription revenue
• Manage admins
• List all admins
• Update admin records
• Delete admin accounts
• Add new admin accounts
• View admin profile
• Update admin profile details
• Change admin password
• Reset admin password
• Manage users, plans, actions, packages, customers, admins, and revenue from MongoDB-backed admin controls

The admin portal includes dashboard, manage plans, manage ML actions, manage packages, manage customers, revenue, all admins, add admin, admin profile, admin password update, and admin password reset modules

• User registration
• User login
• JWT token stored in cookie
• Password reset
• Landing page
• Product overview
• Login and signup links
• User guide page
• User dashboard
• View account status
• Quick links to predictions
• Quick links to data upload
• User panel as central navigation hub
• Profile management
• View name, email, phone, and username
• Update name
• Update email
• Update phone
• Update username
• Change password
• Upload sales CSV
• Upload .csv file only
• Upload file up to 100MB
• Preview uploaded sales data
• Paginated CSV preview
• Delete uploaded data where supported
• Download sample file
• View current subscription plan
• Compare plan tiers
• Upgrade plan through dummy checkout
• Access ML modules based on subscription plan
• Run analysis from prediction pages
• View prediction metrics
• View charts and warnings
• Download result files
• Preview tabular output where available

The user portal includes landing page, login, signup, reset password, user guide, dashboard, user panel, profile, change password, upload sales data, preview data, and plans/billing modules.

Machine Learning Modules

1. Sales Forecasting

• Available from Free plan and above
• Uses Facebook Prophet
• Forecasts daily sales for the next 30 days
• Shows top products
• Shows train/test metrics
• Displays interactive forecast charts

2. Market Basket Analysis

• Available from Standard plan and above
• Uses Apriori algorithm through mlxtend
• Finds product association rules
• Shows support, confidence, and lift
• Requires main sales CSV

3. RFM Segmentation

• Available from Standard plan and above
• Uses K-Means clustering
• Works on Recency, Frequency, and Monetary values
• Segments customers into low, mid, and high value groups

4. Customer Lifetime Value Prediction

• Available from Standard plan and above
• Predicts customer lifetime value
• Uses RFM-style features and clustering

5. Churn Analysis

• Available in Premium plan
• Uses XGBoost classifier
• Predicts customer churn
• Requires separate churn CSV
• Shows accuracy, precision, recall, and confusion matrix

6. Response Modeling

• Available in Premium plan
• Uses XGBoost
• Models marketing response by offer type
• Compares actual vs predicted orders and revenue

7. Uplift Modeling

• Available in Premium plan
• Uses XGBoost and K-Means
• Calculates uplift scores
• Supports campaign ROI analysis
• Identifies persuadables and lost causes

The project provides seven ML modules across plan tiers: sales forecasting, market basket analysis, RFM segmentation, LTV prediction, churn analysis, response modeling, and uplift modeling.

Subscription Plans

Plans and ML action mappings are seeded into MongoDB and can be modified from the Admin Panel.

Other Features

• Full-stack MERN architecture
• React frontend
• Express backend
• MongoDB database
• Python ML scripts
• Node-to-Python integration using python-shell
• CSV upload
• Sales data preview
• Sample CSV download
• Result file generation
• Excel file output storage
• Charts and visualizations
• Confusion matrices
• Forecast graphs
• Data tables
• Subscription-based ML access
• Dummy payment checkout
• Admin revenue management
• User and admin portals
• JWT authentication
• Password reset
• MongoDB-backed user, admin, plan, action, and revenue data
• Static output files served from backend
• Development setup using concurrently
• Demo seed data
• Sample users with Free, Standard, and Premium plans
• Pre-uploaded sales CSV for selected demo accounts
• Python path resolver support

The system architecture shows React client, Express API, Python ML scripts, and MongoDB working together, with uploaded CSV files stored under public/ and ML output files saved under excel_files/{userId}/.

Technology Used

The uploaded documentation lists React, Express, JWT, Multer, MongoDB, Mongoose, Python, pandas, numpy, scikit-learn, Prophet, mlxtend, XGBoost, matplotlib, plotly, python-shell, and concurrently as part of the technology stack.

CSV Requirements

Main Sales CSV

Used by Sales Forecasting, Market Basket Analysis, RFM Segmentation, and LTV Prediction.

Stored as:

public/{userId}.csv

Churn CSV

Required columns include:

customerID, gender, SeniorCitizen, Partner, Dependents, tenure, PhoneService, PaperlessBilling, MonthlyCharges, TotalCharges, Churn

Stored as:

public/{userId}churn.csv

Uplift CSV

Required columns:

history, offer, conversion

Stored as:

public/{userId}uplift.csv

Market Response CSV

Required columns:

offer, conversion

The project documentation defines separate CSV requirements for main sales data, churn analysis, uplift modeling, and market response modeling

Python Scripts Reference

Step 1: Navigate to application folder

cd "Code - Frontend + Backend"

Step 2: Install backend dependencies

npm install

Step 3: Install frontend dependencies

cd client

npm install

cd ..

Step 4: Install Python ML dependencies

pip install prophet pandas numpy scikit-learn mlxtend xgboost openpyxl matplotlib plotly

Step 5: Seed the database

npm run seed

Step 6: Start development servers

npm run dev

This starts both backend and frontend together.

Alternative: Run separately

npm start

npm run client

The setup instructions include installing backend dependencies, frontend dependencies, Python ML packages, seeding the database, and running development servers.

Admin Panel

User App