📃expenses-automation-script v1.0

Python script to automate copy & pasting expenses from unformatted bank statements for accounting.
Explore the docs »

View Demo · Report Bug · Request Feature

---

About The Project

Script to automate copy & pasting expenses from unformatted bank statements for accounting purposes.

This automation project was inspired by the lack of proper xls/csv exporting options from certain banks (e.g. CITI), where we only have a CITI mobile app that provides billing statements in PDF. This was a tedious format to clean and transfer onto spreadsheet programs.

With this script, we can automate the copy & pasting, as well as formating/categorizing of our expenses with ease. This script can be tweaked to serve any other copy/pasting purposes as well.

Written with PyAutoGUI, Pynput, Pyperclip and Natural Language Toolkit (NLTK).

Libraries Used:

• PyAutoGUI - Automating mouse and keyboard actions such as clicking and typing
• Pynput - Event listeners to store variables
• Pyperclip - Clipboard for pasting data
• NLTK - Provides tools to work with human language data. Specifically this script uses a class in the NLTK.classify module, NaiveBayesClassifier, that provides an implementation of the Naive Bayes algorithm for text classification. This machine learning techniques help to predict & classify our expenses category

Release & Updates

• V1.1 - Work in progress
  • Text recognition, planning to remove some of V1.0 's manual tasks
• V1.0
  • Able to copy text from badly formatted PDF files, and paste them onto excel or google sheets via mouse & keyboard actions.
  • some actions are still manual (defining copy & paste coordinates with initial configuration).
  • Food category is still manual (Prediction is not accurate for food related expenses, due to the vast variations of restaurants)

Usage Demo

Usage Guide

1. Clone Repository
2. Open project in your favorite IDE (i.e. VS Code)
3. Install dependencies (libraries)
4. Calibrate the line_spacing field in the Automation class (This represents the y-dist for the cursor to move between transactions in your PDF)
5. Update build_json with your own dataset and expenses, tailor it to fit your own use case
6. Run main.py in the terminal
7. dataset.json & keywords_food.json should be generated in the same directory
8. 1st prompt - select start point for copy script
9. 2nd prompt - select start point for paste script
10. Script executes successfully.

Contact

• https://github.com/nc1z

Project Link: https://github.com/nc1z/expenses-automation-script

(back to top)