Skip to content

AI-Driven Financial Management Platform - Built with TensorFlow for native, on-device predictive analytics. Leveraging native TensorFlow models for automated transaction categorization and budgeting.

Notifications You must be signed in to change notification settings

Satharva2004/WebWallet-AI

Repository files navigation

Insightful Wallets

AI-powered personal finance dashboard and budgeting companion (Vite + React + TypeScript).

Features

  • Auth & profile
  • Dashboard KPIs & charts
  • Transactions CRUD
  • Budgets & savings goals
  • AI insights & advisor chat

Run Commands

npm install
npm run dev

# build
npm run build

# optional: preview build
npm run preview

AI Algorithm Design

Time-series prediction of the future spending patterns is done using an AI insight generation algorithm based on TensorFlow.js. The algorithm gathers the latest 30 days of transaction history, normalizes the values into normalized arrays, and uses a linear regression model to forecast the expenditure in the next period (Zhang and Liu, 2024). The accuracy of prediction increases with increasing availability of historical data, and the system needs at least 10 transactions to give predictions.

The budget analysis algorithm compares the actual expenditure with the set limits, calculates usage percentages, and provides an alert when the expenditure is over 80% of the fixed budgets. The savings rate calculation ratio takes the total savings and the total income and compares the outcome with the recommended saving of 20 percent (Chen, 2024). Financial impact of the insight is given the top priority in the system, where the priority of the budget overruns is high, prediction alerts are given medium priority, and the general tips are given low priority.

About

AI-Driven Financial Management Platform - Built with TensorFlow for native, on-device predictive analytics. Leveraging native TensorFlow models for automated transaction categorization and budgeting.

Topics

Resources

Stars

Watchers

Forks

Releases

No releases published

Packages

No packages published

Languages