Cookie Warrior : Product Intelligence Workspace

Presentation File

View the Google Slides Presentation

The Live Demo

Experience the live application here: UniHack - Product Intelligence · Streamlit

---

The Overview

Welcome to the MVP repository for the UniHack Hackathon! This project provides a robust data enrichment pipeline designed to solve incomplete Product Information Management (PIM) data.

By leveraging automation in app_final.py, the pipeline ingests minimal product identifiers and aggressively enriches them, generating everything from marketing descriptions and SEO-friendly metadata to granular technical attributes (e.g., voltages, dimensions, sound levels) and media URLs.

Repository Structure

As seen in image_51a3fe.png, the core of the repository consists of:

• app_final.py: The main application Streamlit script containing the pipeline logic.
• requirements.txt: The list of Python dependencies required to run the pipeline.
• README.md: Project documentation (this file).

Dataset Reference

To understand the pipeline's transformation capabilities, we have provided two reference files:

1. The Input

File: Unihack_ Sample Dataset - Input (1).csv This file represents the raw data starting point. It contains only 6 basic columns:

• Mfg_Part_Num
• Part_Desc
• E1_Brand
• Unilog_Brand
• DIB_Brand
• Part_Manuf

2. The Output

File: Unihack_ Expected Output - Delivery Format (1).csv This file demonstrates the final, enriched delivery format. The pipeline maps the 6 input columns into a massive schema of over 100 columns, including:

• Categorization: Dept, Class, Fine, UNSPSC.
• Rich Descriptions: MOBILE_DESC, INVOICE_DESC, SHORT_DESC, MARKETING_DESCRIPTION.
• Extracted Features: Up to 20 individual ITEM_FEATURES (e.g., "Adjustable 2nd Rack", "Leak Detection System").
• Dynamic Attributes: 50 dynamically mapped ATTRIBUTE_LABEL, ATTRIBUTE_VALUE, and ATTRIBUTE_UOM (Unit of Measure) fields for precise technical specifications.
• Digital Assets: Image URLs, MFR URL, Specification Sheets, and Warranty documents.

Installation & Setup

1. Clone the repository:

   git clone [https://github.com/KevinJonathan7708/unihack-hackathon-mvp.git](https://github.com/KevinJonathan7708/unihack-hackathon-mvp.git)
   cd unihack-hackathon-mvp

2. Install the dependencies:

   pip install -r requirements.txt

3. Run the application locally:

   streamlit run app_final.py

Built With

• Python 3
• Streamlit (for the web interface)
• Pandas (for complex data transformations)

💡 Use Case

This MVP is built to automate away the hundreds of manual hours typically required by e-commerce managers and data stewards to clean, standardize, and enrich product catalogs before going to market.

##App Screenshot