ReceiptAnalysis

Project for a receipt analysis of a dataset from kaggle.

Full report (CS only) is here

Requirements

• This project has not been optimized to run on a PC with any RAM size, 32Gb is thus recommended to analyze 100k of records.
• For more records, optimization or bigger RAM is necessary.
• Python v3.11.x (used through pyenv)

Usage

1. Clone the repository
2. Register to Kaggle
3. Download dataset eCommerce purchase history from electronics store
4. Place the dataset (kz.csv) into /data directory
5. Install poetry
6. Install depencences cd /path/to/ReceiptAnalysis && poetry install
7. Run preprocessing poetry run preprocessing
8. Run clustering poetry run clustering
9. Generate associative rules poetry run associative_rules

Outputs

• Saved numpy matrices are in /data
• Saved matplotlib figures are in /images
• Saved text outputs of the scripts are in /outputs
• Saved clusters and cluster rules are in /rules

Used tutorials

• Black, MyPy and Pylint
• Pyenv and poetry