Dataset¶
This project starts from the public Nova 50M Transactions Super-App dataset, a synthetic marketplace dataset representing transactions, users, services, and operating markets for a global super-app.
What It Contains¶
The source data includes four core entities:
| Entity | What it represents | Example fields |
|---|---|---|
| Interactions | Transaction-level marketplace activity | User, service, category, amount, status, timestamp, platform, referral source, location |
| Users | Customer profile attributes | Membership tier, device, acquisition source, market |
| Services | Service and category metadata | Category, rating, marketplace availability |
| Markets | City-level operating markets | Market name, region, latitude, longitude |
How We Used It¶
We treated the original dataset as a starting point, then regenerated the dataset for realism before building the final analytics layer.
The regeneration step preserved the project scale and core schema, but added more realistic business behavior: market-level seasonality, category-specific demand and basket sizes, weather sensitivity, service supply effects, customer lifecycle variation, and fraud-like transaction patterns.
The dashboards and dbt models in this project use the regenerated dataset, not the raw Kaggle files directly.