license: apache-2.0
task_categories:
- text-classification
- feature-extraction
language:
- en
tags:
- finance
- stock rapport
- northwind
- company documents
Northwind Stock Report Dataset
This dataset was created by CHERGUELAINE Ayoub & BOUBEKRI Faycal for the purpose of document classification and analytics. The dataset contains monthly stock reports and monthly stock reports by category, extracted from the Northwind dataset.
The Northwind dataset is a sample database that comes with Microsoft Access, and is commonly used as a demo database for learning SQL. The dataset contains data on a fictional company called "Northwind Traders", which imports and exports specialty foods.
To create the dataset of stock reports, we used SQL queries to extract data from the Northwind database, and then converted the results to PDF documents using the ReportLab library in Python. The resulting PDFs were then manually labeled with their corresponding categories.
We used ChatGPT, a language model trained by OpenAI, to help us write the Python scripts to extract the data and generate the PDFs. ChatGPT provided us with valuable guidance and assistance throughout the project.
Dataset Structure
The dataset contains two subdirectories: monthly_reports and category_reports.
Monthly Reports
The monthly_reports directory contains monthly stock reports, where each file represents a stock report for a specific month. The filenames are in the format YYYY_MM.pdf, where YYYY is the year and MM is the month.
Category Reports
The category_reports directory contains monthly stock reports by category, where each file represents a stock report for a specific month and category. The filenames are in the format YYYY_MM_Category.pdf, where YYYY is the year, MM is the month, and Category is the name of the category.
Acknowledgments
We would like to thank ChatGPT, a language model trained by OpenAI, for providing valuable guidance and assistance throughout the project.