dream-machine-ai/Dream-Machine-08-09-Dataset

██████╗  █████╗ ████████╗ █████╗ ███████╗███████╗████████╗
██╔══██╗██╔══██╗╚══██╔══╝██╔══██╗██╔════╝██╔════╝╚══██╔══╝
██║  ██║███████║   ██║   ███████║███████╗█████╗     ██║
██║  ██║██╔══██║   ██║   ██╔══██║╚════██║██╔══╝     ██║
██████╔╝██║  ██║   ██║   ██║  ██║███████║███████╗   ██║
╚═════╝ ╚═╝  ╚═╝   ╚═╝   ╚═╝  ╚═╝╚══════╝╚══════╝   ╚═╝

Dream-Machine · E-Commerce Synthetic Dataset

1,000 Products · 1,000 Users · 5,000 Behaviors · JD/Taobao Schema · Fully Reproducible

"Fully synthetic, fully reproducible — from 1K to 10B items, the schema stays the same."

📦 Dataset Files · 🔍 Schema · 🔄 Reproduce · 🤗 Model · ⚖️ License

---

✨ What Makes This Dataset Different

✦ JD/Taobao schema   — mirrors real-world e-commerce data structures
✦ Fully synthetic    — zero real PII, safe for open research
✦ Fully reproducible — three Python scripts regenerate everything from scratch
✦ Multi-level        — products, users, and interaction behaviors in one package
✦ Scale-ready        — schema supports 10² → 10¹⁰ item catalogues unchanged

---

📦 Dataset Contents

Directory	Files	Records	Size	Description
data/products/	5 JSON	1,000	~2.9 MB	Product catalogue (JD/Taobao schema)
data/users/	5 JSON	1,000	~5.0 MB	User profiles with demographics & interests
data/behaviors/	5 JSON	5,000	~3.4 MB	Interaction logs: view / cart / purchase / rate
scripts/	3 Python	—	~50 KB	Fully reproducible generation scripts

All data is batched into 5 files per split (200 records/batch for products & users; 1,000 records/batch for behaviors), making it easy to stream or load incrementally.

---

🔍 Data Schema

Product (data/products/products_batch_*.json)

{
  "sku_id":   "SKU000001",
  "spu_id":   "SPU000001",
  "title":    "Nike Air Jordan 1 Retro High OG",
  "subtitle": "经典复刻，限量发售",
  "category": {
    "level1_id":   "C01",
    "level1_name": "运动",
    "level2_id":   "C0102",
    "level2_name": "鞋靴",
    "level3_id":   "C010201",
    "level3_name": "篮球鞋"
  },
  "brand": {
    "brand_id":   "B001",
    "brand_name": "Nike"
  },
  "price": {
    "current_price":  1499.0,
    "original_price": 1999.0,
    "discount":       7.5
  },
  "sales": {
    "total_sales":   320000,
    "monthly_sales": 18000
  },
  "rating": {
    "overall":      4.95,
    "quality":      4.97,
    "review_count": 12800
  },
  "rec_features": {
    "ctr_7d":    0.21,
    "cvr_7d":    0.09,
    "like_rate": 0.88
  }
}

User (data/users/users_batch_*.json)

{
  "user_id": "U000001",
  "demographics": {
    "gender":       "female",
    "age_group":    "25-30",
    "income_level": "10000-20000",
    "occupation":   "白领"
  },
  "location": {
    "city":      "北京",
    "city_tier": "一线",
    "province":  "北京市"
  },
  "interests": {
    "interest_categories": ["数码科技", "运动健身"],
    "brand_preferences":   ["Nike", "Apple"],
    "style_tags":          ["简约", "运动"]
  },
  "consumption": {
    "price_sensitivity": "medium",
    "avg_order_value":   350.0,
    "purchase_freq":     "monthly"
  },
  "activity": {
    "active_days_30d": 22,
    "last_active":     "2026-08-19"
  }
}

Behavior (data/behaviors/behaviors_batch_*.json)

{
  "behavior_id":  "B000001",
  "user_id":      "U000001",
  "sku_id":       "SKU000042",
  "action":       "purchase",
  "timestamp":    "2026-07-15T14:32:00",
  "dwell_time_s": 120,
  "rating":       5,
  "scene":        "homepage",
  "device":       "mobile"
}

Action types:

Action	Description
view	Product detail page view
cart	Added to shopping cart
purchase	Completed purchase
rate	Left a review / rating
like	Saved to favourites

---

🔄 Reproduce the Data

All data is fully reproducible from the included Python scripts. The scripts have no external dependencies beyond numpy and faker.

# Install dependencies
pip install numpy faker

# Regenerate all data (outputs to data/)
python scripts/generate_product_data.py   # → data/products/
python scripts/generate_user_data.py      # → data/users/
python scripts/generate_behavior_data.py  # → data/behaviors/

Set SEED=42 (default) for deterministic output. Change the seed to generate alternative splits.

---

📊 Statistics

Feature	Products	Users	Behaviors
Total records	1,000	1,000	5,000
Category levels	3 (L1/L2/L3)	—	—
Unique categories (L1)	12	—	—
Unique brands	50	—	—
City tiers	—	5 (一线→五线)	—
Age groups	—	8	—
Action types	—	—	5
Date range	—	—	2026-01 ~ 2026-08

---

🤗 Related Resources

Resource	Link
Model	dream-machine-ai/Dream-Machine-08-09
Paper	Magic Shop at Scale (Zenodo)
GitHub	Dream-Machine

---

📖 Citation

If you use this dataset in your research, please cite the accompanying paper:

@techreport{wen2026dreammachine,
  title       = {Magic Shop at Scale: Dream-Machine --- A Billion-Item
                 Real-Time Neural Recommendation Sorcerer},
  author      = {Wen, Fangjun},
  year        = {2026},
  institution = {DreamMachine Research Team},
  doi         = {10.5281/zenodo.21906715},
  url         = {https://doi.org/10.5281/zenodo.21906715},
  note        = {Technical Report, August 2026. arXiv cs.IR / cs.LG}
}

---

⚖️ License

This dataset is licensed under the Creative Commons Attribution-NonCommercial 4.0 International License (CC BY-NC 4.0).

✅ Permitted	❌ Prohibited
Academic research & publication	Commercial products or services
Personal learning & experimentation	Revenue-generating deployments
Non-commercial derivative works	Sublicensing for profit
Citing in papers with attribution	Any business use without written permission

All data is fully synthetic. It contains no real personal information, real user data, or real commercial transaction data. Any resemblance to real persons, products, or businesses is purely coincidental.

Commercial use of any kind requires explicit prior written consent from the author (Fangjun Wen). Contact: fangjunwen168@outlook.com

See LICENSE for the full license text.

---

Paper: Magic Shop at Scale: Dream-Machine · Authors: Fangjun Wen · DreamMachine Research Team · August 2026

// built with focus · DreamMachine Research