Dataset — Olist Brazilian E-commerce
What it is
Real anonymized data from Olist, a Brazilian e-commerce marketplace. ~100,000 orders placed between 2016 and 2018, with customers, sellers, products, payments, and customer reviews.
Source: Kaggle — olistbr/brazilian-ecommerce License: CC BY-NC-SA 4.0 (non-commercial, share-alike) Size: ~135 MB unzipped
Why we chose it
- 8 related tables — makes Day 2 (SQL joins) meaningful instead of toy
- ~100K orders — fits in Excel (well under the 1M row limit)
- Free-text reviews — gives Day 5 (Claude Code) something genuinely AI-shaped to do
- Real business problems — delivery delays, payment mix, seller performance, review drivers
- Well-documented schema — easy to onboard students
Table overview
| Table | Rows (approx) | What it contains |
|---|---|---|
olist_orders_dataset | 99,441 | One row per order: status, purchase/delivery timestamps |
olist_order_items_dataset | 112,650 | One row per line item: product, seller, price, freight |
olist_order_payments_dataset | 103,886 | Payment method, installments, value |
olist_order_reviews_dataset | 99,224 | Review score (1–5) and free-text comments |
olist_customers_dataset | 99,441 | Customer location (state, city, zip prefix) |
olist_sellers_dataset | 3,095 | Seller location |
olist_products_dataset | 32,951 | Product category, dimensions, weight |
product_category_name_translation | 71 | Portuguese → English category names |
Schema diagram is in the Kaggle page linked above.
How to get the data
Step 1 — download the CSVs
bash data/olist/download.shThis downloads the Kaggle zip and unpacks all CSVs into data/olist/. The CSV files themselves are gitignored — every student downloads them locally.
If the kaggle CLI is not set up, the script prints manual download instructions.
Step 2 — load into SQLite (do this before Day 2)
Day 2 (SQL) uses SQLite. Build a single olist.db database file from the CSVs in one command:
cd data/olist
sqlite3 olist.db < load_into_sqlite.sqlYou should see row counts print at the end (~99K orders, ~112K items, etc.). If you see “Error: no such file” for any CSV, re-run the download in Step 1.
After this you have data/olist/olist.db — that’s the database used by every Day 2 exercise and the SQL capstone. Open it with:
- The CLI:
sqlite3 data/olist/olist.db - A GUI (recommended for beginners): DB Browser for SQLite — free, runs on Mac/Windows/Linux, looks like Excel for databases.