AppLink: [Mpesa Finance App]

The M-Pesa Finance App is an AI-powered personal finance assistant built to clean your M-Pesa statements, analyze transactions, detect spending patterns, and forecast future expenses. Think of it as your budget’s watchdog — but smarter and way less dramatic.

• Automated Data Extraction

1. Upload your M-Pesa PDFs
2. The app parses → cleans → structures your data
3. Removes footers, line breaks, duplicated spaces, inconsistent fields

• Data Preprocessing & Cleaning

1. Handles messy formats and weird characters (like Withdraw\r\n)
2. Converts dates, balances, and amounts correctly
3. Detects outliers using IQR
4. Normalizes and standardizes transaction categories

• Interactive Analysis Dashboard

1. Total expenses, income, balances
2. Category-wise insights (Send Money, Buy Goods, PayBill, Withdraw, etc.)
3. Time filters (daily, weekly, monthly)
4. Dynamic trend visualizations

• Forecasting Engine (AI + Time Series Models)

1. Uses Prophet / ARIMA to forecast expenses
2. Evaluates predictions using RMSE and MAE
3. Supports future multivariate forecasting
4. Predicts upcoming spending trends

• Smart Alerts & Insights

1. Flags suspicious or unusual spending
2. Notifies about overspending streaks
3. Visualize your spendings

• Image & Report Saving

1. Export charts and analysis screenshots directly
2. Useful for financial tracking or monthly reports

1. Clone the repository

git clone https://github.com/Brian342/MpesaFinance
cd MpesaApp.py

2. Install dependencies

pip install -r requirements.txt

4. Run the app

streamlit run MpesaApp.py

Python

Streamlit

Pandas / NumPy

Prophet / ARIMA

Matplotlib / Plotly

Django (future API)

This project is licensed under the MIT License. See the LICENSE file for more details.