SpendWise AI – Executive Financial

SpendWise AI – Executive Financial Audit Dashboard

📌 Project Overview

SpendWise AI is an executive-grade financial intelligence dashboard designed to ingest multi-currency transactional logs and deliver automated data auditing, dynamic runway projections, and algorithmic risk flags. Built with a responsive dark-themed corporate UI, this application transforms raw transaction histories into an interactive compliance and financial health management portal.

🚀 Key Engineering & Analytics Features

Multi-Currency Vectorized Financial Metrics: Separates and aggregates transaction lists based on sign conventions using vectorized arrays. It computes real-time total income, total expenses, and net balance under flexible, user-selected global currencies (including PKR, USD, and GBP).

Predictive Financial Runway Engine: Dynamically calculates an individual or business's average daily burn rate by calculating the elapsed delta between the earliest and latest transaction timestamps. If a dataset lacks chronological entries, a robust safety fallback algorithm takes over, shifting to an adaptive 30-day timeline simulation to maintain zero application runtime failures.

Algorithmic Outlier & Anomalous Spending Audit: Employs a dynamic rule-based statistical thresholding mechanism ($3\times$ the calculated mean expense value). Any transaction breaching this boundary is caught by an automated inspection pipeline, which marks the record with an operational warning (🚩 RED FLAG: High Spend) and isolates it in a standalone high-risk ledger for immediate review.

Production-Grade Data Resiliency: Engineered with an assertive preprocessing pipeline including automatic column trim operations (df.columns.str.strip()) and strict numeric enforcement (pd.to_numeric with errors='coerce'). This ensures the app safely processes poorly formatted, manually generated, or commercial banking files without crashing.

🛠️ Technology Stack

User Interface: Streamlit (Python-powered, data-reactive framework)

Data Processing Layer: Pandas & NumPy (Vector mathematics, database filtering, and string normalization)

Custom Styling: Embedded custom CSS styling overrides for deep-tint metric elements