I started this project to help a self employed friend to do his tax return. Since his expenses are very repetitive, instead of manually selecting the right category, I thought developing a piece of code that learns from the previous years categories would make the whole process much faster.
This script was developed for a Nationwide bank statements, but can easily adapted to any statement, you just need to make some changes in the lines reading the csv. The categories will depend on the type of business you run. It will be customised for your specific needs based on the 'expenses_train.csv' file that the code reads.
For privacy reasons, I do not share any of the files used to train and test the model since they all contain personal and indentifiable information.
If you would like to use this piece of code and have any questions please do not hesitate to get in touch with me (camino.bruno@gmail.com).