Data - Olist Brazilian E-commerce

The dataset is not committed to the repo. It's ~45 MB across 9 CSV files. Get it once and the ETL pipeline takes care of everything from there.

Acquire the data

Option A - Kaggle CLI (preferred, scriptable)

# 1. Set up Kaggle credentials once: https://www.kaggle.com/docs/api
pip install kaggle
# Place kaggle.json at ~/.kaggle/kaggle.json (chmod 600)

# 2. Download into data/raw/
cd data/raw/
kaggle datasets download -d olistbr/brazilian-ecommerce --unzip

Option B - Manual download

1. Go to https://www.kaggle.com/datasets/olistbr/brazilian-ecommerce
2. Click Download (requires a free Kaggle account)
3. Unzip the contents directly into data/raw/

Expected files

After unzipping you should have (filenames are exact):

data/raw/
├── olist_customers_dataset.csv
├── olist_geolocation_dataset.csv
├── olist_order_items_dataset.csv
├── olist_order_payments_dataset.csv
├── olist_order_reviews_dataset.csv
├── olist_orders_dataset.csv
├── olist_products_dataset.csv
├── olist_sellers_dataset.csv
└── product_category_name_translation.csv

Schema overview

Table	Rows (≈)	Grain	Key joins
orders	99,441	one per order	customer_id → customers; order_id → items, payments, reviews
order_items	112,650	one per order × seller × product	order_id, product_id, seller_id
order_payments	103,886	one per payment instalment	order_id
order_reviews	99,224	one per review	order_id
customers	99,441	one per customer-order shipping address	customer_unique_id is the dedup key
products	32,951	one per product	product_id, product_category_name
sellers	3,095	one per seller	seller_id, seller_zip_code_prefix
geolocation	1,000,163	one per zip prefix × lat/lng	many duplicates per zip - needs dedup
category_translation	71	category name PT → EN	product_category_name

A note on grain: customers has one row per (customer × shipping address × order), so customer_id is not a stable user ID across orders - customer_unique_id is. This is the kind of subtlety the silver layer cleans up.

License

Olist Brazilian E-commerce Public Dataset © Olist Store, released under CC BY-NC-SA 4.0. Non-commercial use only.