Every file in this course is real, public data: no synthetic rows, no edited values. Each lab and homework carries its files in data/raw/, committed on purpose and never edited, with a DATA.md beside them giving the exact source, the date it was fetched, its license and attribution, and what was changed for the course (a subset, a format). The repository’s LICENSE.md lists each dataset’s license. This page is the catalogue.
The names of places are the publishers’. The country names, area names and codes in these files are the ones their publishers use. Their use in this course says nothing about the political status of any place. In the labs, country is the file’s word: see the glossary.
What a file does — which rows are not what they claim to be, what “missing” looks like in it — is not written here. Finding it is the work, and the census is how.
Online Retail II
What it is
Every invoice line of a UK-based online retailer of gifts, 1 December 2009 to 9 December 2010.
Source
Daqing Chen, Online Retail II, UCI Machine Learning Repository, doi:10.24432/C5CG6D (the 2009–2010 sheet).
License
CC BY 4.0.
As shipped
online_retail.parquet: all 525,461 lines of that sheet, 8 columns, 3.2 MB. Converted from CSV to Parquet with the types DuckDB inferred; no rows or values changed.
Used in
the setup check; the Block 1 lecture (in the Lab 1 project); the Block 2 lecture’s “why SQL” demonstration; Lab 2; Homework 1; the final exam and the mock.
Our World in Data — CO₂ and greenhouse-gas emissions
What it is
Emissions, population, and GDP by country and year, as Our World in Data publishes them.
CC BY 4.0 for Our World in Data’s own work; each underlying series keeps its source’s terms, named per column in the codebook. Cite Our World in Data and the named source.
As shipped
owid_co2.csv: 15 of the file’s columns, years 1990 to 2024, 8,883 rows; owid_codebook.csv, what each of those columns means. countries.csv (7,665 rows) is a further subset, made with the filter Lab 1 builds, for Block 2.
Used in
Lab 1; the Block 2 lecture (countries.csv, in the Lab 2 project); the final exam and the mock.
Brazilian E-Commerce Public Dataset by Olist
What it is
Orders from a Brazilian online marketplace, 2016 to 2018, in eight linked tables: orders, their items, payments, and reviews; customers, products, sellers; and a translation of the product categories.
CC BY-NC-SA 4.0. NonCommercial: redistributed here for teaching only. ShareAlike: the subset is an adaptation, shared under the same license. The course’s own code and text are not relicensed by it.
As shipped
A selected subset, not a representative one: 2,653 of the 99,441 orders, chosen so that the situations the labs are about are present, with every item, payment, and review of those orders and every customer, product, and seller they refer to. The selection rule is in DATA.md. orders 2,653 rows · order_items 3,024 · order_payments 2,790 · order_reviews 2,681 · customers 2,653 · products 2,217 · sellers 903 · product_category_name_translation 71.
Used in
the Block 3 and Block 4 lectures; Lab 3 (items, products and category names); Lab 4 (orders and payments); Homework 2 (orders, payments, customers and reviews); Knowledge check 3 and its ungraded retry, the final exam and the mock, which quote small, unchanged extracts.
Eurostat — municipal waste (env_wasmun)
What it is
Municipal waste by waste-management operation, country, and year, in Eurostat’s own file layout.
Source
Eurostat, dataset env_wasmun, through Eurostat’s dissemination API (TSV).
env_wasmun.tsv, unmodified: 742 rows, one column per year from 2010 to 2024. The first column packs four codes into one; Eurostat’s code lists for them and for its observation flags ship beside it (codelist_*.tsv).
Used in
Lab 6 (waste per person); Homework 3 (the waste totals, in thousand tonnes); the final exam.
Frankfurter — European Central Bank reference rates
What it is
The ECB’s daily euro reference rates for the Brazilian real and the US dollar, 1 September 2016 to 31 October 2018, as one API response.
Source
The Frankfurter API, which serves the European Central Bank’s euro foreign exchange reference rates. The request is written in DATA.md.
License
ECB statistics may be reused with the source acknowledged: Source: European Central Bank. The Frankfurter API is open source (MIT).
As shipped
api/frankfurter_eur.json: the response exactly as received — one JSON document with rates for 555 dates. The ECB publishes on its business days only.
Used in
No lab or homework this year: Session 3 reads the World Bank’s pages instead.
World Bank — World Development Indicators
What it is
Population (SP.POP.TOTL) and GDP at purchasing-power parity in constant international dollars (NY.GDP.MKTP.PP.KD), all countries and groups of countries, 2010 to 2024; and the World Bank’s list of countries and groups.
Source
World Bank Indicators API v2, api.worldbank.org. The requests are written in DATA.md.
License
CC BY 4.0. Attribution: World Bank, World Development Indicators.
As shipped
Each API response exactly as received, one file per page, under api/worldbank/: four pages per indicator (3,975 records each), and one page of country metadata (295 entries).
Used in
The Block 5 lecture (GDP) and Lab 5 (population, the country list), in the Lab 5 project; the Block 6 lecture (GDP and population, in the Lab 6 project); Homework 3 (population); the final exam and the mock.
EU-27 country codes
What it is
The 27 EU member states (as of 2020), one row each, with the code each source uses for them: Eurostat’s, the ISO three-letter code Our World in Data uses, and the World Bank’s.
Source
Made for this course, by hand, from the three sources’ own documentation.
License
CC BY 4.0, as the course materials.
As shipped
country_codes.csv, 27 rows.
Used in
Homework 3.
Using the data outside the course
Follow each dataset’s license. For Olist that means non-commercial use only, and the same license on anything you share that is built from it. For the others, attribution as written above. Your homework zip contains data/raw/, DATA.md, and LICENSE.md: it is for grading only.