brijeshvadi/eprocure-product-catalog

---
license: cc-by-nc-4.0
language:
- en
- ar
tags:
- procurement
- b2b
- e-commerce
- bilingual
- saudi-arabia
- product-catalog
size_categories:
- 1K<n<10K
task_categories:
- text-classification
- sentence-similarity
configs:
- config_name: default
data_files:
- split: train
path: data/products.csv
- split: validation
path: data/rfq_pairs.csv
---

# e-Procure Product Catalog

Bilingual (English/Arabic) B2B procurement product catalog from the Saudi Arabian construction and industrial supply chain.

## Dataset Description

Contains 2,400 product listings across 5 categories with English and Arabic descriptions, pricing tiers, and supplier metadata. Paired with 800 RFQ (Request for Quotation) matching pairs for training similarity models.

### Categories

| Category | Products | Suppliers | Avg Price (SAR) |
|----------|----------|-----------|-----------------|
| Construction Materials | 680 | 24 | 4,250 |
| Electrical Equipment | 540 | 18 | 8,900 |
| HVAC Systems | 420 | 12 | 15,600 |
| Plumbing Supplies | 380 | 15 | 2,100 |
| Safety Equipment | 380 | 20 | 850 |

## Schema — products.csv

| Column | Description |
|--------|-------------|
| `sku` | Unique product SKU |
| `name_en` | English product name |
| `name_ar` | Arabic product name |
| `description_en` | English description |
| `description_ar` | Arabic description |
| `category` | Product category |
| `subcategory` | Product subcategory |
| `brand` | Manufacturer brand |
| `price_sar` | Price in SAR |
| `unit` | Unit of measure |
| `supplier` | Supplier company name |
| `lead_time_days` | Delivery lead time |
| `min_order_qty` | Minimum order quantity |
| `certifications` | Quality certifications |

## Usage

```python
from datasets import load_dataset

ds = load_dataset("brijeshvadi/eprocure-product-catalog")
products = ds["train"]
rfq_pairs = ds["validation"]

electrical = products.filter(lambda x: x["category"] == "Electrical Equipment")
```