Free data sets for maths IAs and EPQs

How to use a source well

• Download from the primary source (the organisation that collected the data) where you can, and cite it in a footnote where you use it and in your bibliography.
• Use an Our World in Data chart to see the pattern first. Cite it as: Our World in Data, ‘<chart title>’, CC BY 4.0. Original source: <the provider the chart names>.
• Read the licence. Most sources here are free to reuse with credit; WHO data is for non-commercial use only, and where no reuse licence exists we say so: copy only the values you analyse and cite the page.

The ranked list

1. Our World in Data Population and health
   Thousands of charts on population, health, energy, climate and the economy. Every chart page has a table view, names its original sources and has a CSV download.

   Format: Charts, CSV · Licence: CC BY 4.0 for charts and data made by Our World in Data; series from other providers keep their own licence (shown on each chart). · Sign-up: None

   IA ideas that use it: Are NBA players getting taller? · Do more doctors mean fewer infant deaths? · Does Benford's law hold for the populations of the world's towns? · Does money buy a longer life? GDP and life expectancy · Exponential or logistic? Modelling my country's population · How fast did COVID-19 spread at the start? Comparing doubling times and 10 more.

2. World Bank Open Data Population and health
   Development indicators for every country, mostly from 1960: population, GDP, life expectancy, infant mortality, immunisation.

   Format: CSV, Excel · Licence: CC BY 4.0 for most datasets (the World Bank's licence page lists any exceptions). · Sign-up: None

   IA ideas that use it: Do more doctors mean fewer infant deaths? · Does Benford's law hold for the populations of the world's towns? · Does money buy a longer life? GDP and life expectancy · Exponential or logistic? Modelling my country's population · How much vaccination is enough to stop measles? · Measuring inequality: fitting Lorenz curves and computing the Gini coefficient.

3. Office for National Statistics (UK) Money and prices
   UK population estimates, births, earnings and the CPI and CPIH inflation series.

   Format: CSV, Excel · Licence: Open Government Licence v3.0: free to reuse with an acknowledgement. · Sign-up: None

   IA ideas that use it: Are elite footballers more likely to be born early in the selection year? · Has my pocket money kept up with inflation? · How much does it cost to start saving ten years later? · Is a longer mortgage ever the better deal? · Should bigger hospitals get bigger regions? Weighted Voronoi diagrams · Should I lease or buy my first car? and 4 more.

4. US births per day (NCHS and SSA, via FiveThirtyEight) Population and health
   Births on every day of the year in the US, 1994–2014, from the National Center for Health Statistics and the Social Security Administration.

   Format: CSV · Licence: CC BY 4.0 (FiveThirtyEight's data repository); cite NCHS and SSA as the original sources. · Sign-up: None

   IA ideas that use it: Are elite footballers more likely to be born early in the selection year? · The birthday problem, extended: what about birthdays within a day?.

5. UN World Population Prospects Population and health
   UN population estimates from 1950 and projections to 2100, by country, age and sex.

   Format: CSV, Excel · Licence: Check the terms on the UN site. · Sign-up: None

   IA ideas that use it: Does Benford's law hold for the populations of the world's towns? · Exponential or logistic? Modelling my country's population · Projecting a country's age structure with a Leslie matrix.

6. NASA GISTEMP v4 Climate and environment
   Global and hemispheric temperature anomalies from 1880, monthly and yearly.

   Format: CSV, text · Licence: US government data; NASA asks you to cite the dataset. · Sign-up: None

   IA ideas that use it: How strongly is global temperature linked to ln(CO₂)?.

7. NOAA Global Monitoring Laboratory: Mauna Loa CO₂ Climate and environment
   Monthly and yearly mean atmospheric CO₂ since 1958 (the Keeling curve).

   Format: CSV, text · Licence: Free to use; cite the dataset, its version and the date you downloaded it. · Sign-up: None

   IA ideas that use it: How strongly is global temperature linked to ln(CO₂)? · Modelling the Keeling curve: trend plus seasons.

8. Met Office historic station data Climate and environment
   Monthly maximum and minimum temperature, rainfall, frost days and sunshine for 37 UK stations, some from 1853.

   Format: Text tables · Licence: Check the terms on the Met Office page. · Sign-up: None

   IA ideas that use it: Are our summers getting wetter or drier? · How much does a cold month cost my family? · How much does rain reduce bike-share use?.

9. Gapminder Population and health
   Long-run country indicators (income, life expectancy, child mortality).

   Format: CSV, Excel · Licence: CC BY 4.0. · Sign-up: None

   IA ideas that use it: Does money buy a longer life? GDP and life expectancy.

10. Global Forest Watch Climate and environment
    Yearly tree-cover loss by country since 2001, from satellite imagery.

    Format: CSV, maps · Licence: CC BY 4.0 for the tree-cover loss data. · Sign-up: None

    IA ideas that use it: Is deforestation in [a country] slowing down?.

11. NSIDC Sea Ice Index Climate and environment
    Daily and monthly Arctic and Antarctic sea-ice extent since 1978.

    Format: CSV · Licence: Free to use; NSIDC asks you to cite the dataset. · Sign-up: None for the index files

    IA ideas that use it: When might the Arctic be nearly ice-free in September?.

12. USGS Earthquake Catalog Climate and environment
    Every recorded earthquake with time, place, depth and magnitude; filter and export.

    Format: CSV, GeoJSON · Licence: Check the terms on the USGS site. · Sign-up: None

    IA ideas that use it: How often do big earthquakes happen? The Gutenberg–Richter law.

13. NOAA National Hurricane Center (HURDAT2) Climate and environment
    Track, wind and pressure every six hours for every Atlantic tropical cyclone since 1851.

    Format: Text · Licence: US government data; cite the dataset. · Sign-up: None

    IA ideas that use it: Are Atlantic hurricanes becoming more frequent?.

14. data.police.uk Population and health
    Street-level crimes by month for England, Wales and Northern Ireland.

    Format: CSV · Licence: Open Government Licence v3.0. · Sign-up: None

    IA ideas that use it: Is crime seasonal? Testing monthly crime counts.

15. HM Land Registry Price Paid Data Money and prices
    Every residential sale in England and Wales since 1995.

    Format: CSV · Licence: Open Government Licence v3.0 (read the address-data note on the page). · Sign-up: None

    IA ideas that use it: How much is being near a station worth? · Is a longer mortgage ever the better deal?.

16. GOV.UK rates and thresholds Money and prices
    Current UK income tax, National Insurance and student loan repayment rates, and the Highway Code's stopping distances.

    Format: Web pages · Licence: Open Government Licence v3.0. · Sign-up: None

    IA ideas that use it: Why do stopping distances grow so fast? Modelling the Highway Code table · Will I ever repay my student loan?.

17. ECB Data Portal Money and prices
    Daily euro reference exchange rates and ECB interest rates.

    Format: CSV · Licence: Free to reuse with the source acknowledged. · Sign-up: None

    IA ideas that use it: Is the euro–dollar exchange rate a random walk?.

18. Transport for London open data Maps and travel
    Station entry and exit counts, cycle-hire journeys, and live and timetabled bus arrivals for London.

    Format: CSV, API · Licence: Check TfL's open data terms; attribution is required. · Sign-up: None for the downloads; a free key for heavy API use

    IA ideas that use it: How much does rain reduce bike-share use? · Which Underground station is the most central? · Why do buses come in bunches? A simulation of a bus route.

19. OpenStreetMap Maps and travel
    A free world map with exact coordinates of schools, hospitals, shops, stations, lakes and coastlines.

    Format: Map, export · Licence: Open Database Licence: credit “© OpenStreetMap contributors”; data you build from it must be shared alike. · Sign-up: None to view and read off coordinates

    IA ideas that use it: A fractal shortcut for the travelling salesman: ordering stops along a Hilbert curve · Balancing traffic flow through a road network with linear equations · Do school catchment areas match the Voronoi diagram of the schools? · How big is my local lake? Estimating area with the trapezoidal rule · How far can you see from the top of a hill? · How long is a coastline? Measuring fractal dimension and 14 more.

20. OpenFlights data Maps and travel
    Airports with latitude and longitude, and airline routes.

    Format: CSV · Licence: Open Database Licence. The routes file is a 2014 snapshot. · Sign-up: None

    IA ideas that use it: How many flights does it take to get anywhere? · Which airport is really closest? Voronoi diagrams on a curved Earth · Why do flight paths curve on the map?.

21. Football-Data.co.uk Sport and games
    Results, shots, cards and bookmaker odds for 25+ seasons of European leagues.

    Format: CSV · Licence: No reuse licence found: link to it and cite it, and do not republish the files. · Sign-up: None

    IA ideas that use it: Can a simple rating system predict football results? · Does the best team win? Knockout cups versus leagues · Is home advantage in football disappearing? · Ranking teams that haven't all played each other: the Colley matrix.

22. Lichess open database Sport and games
    Every rated game played on Lichess, in monthly files (very large; start with an early month).

    Format: PGN · Licence: Check the terms on the page. · Sign-up: None

    IA ideas that use it: How big is White's advantage in chess, and does it depend on skill?.

23. World Athletics records and toplists Sport and games
    All-time lists, records and record progressions for every athletics event.

    Format: Web tables · Licence: No reuse licence found: link and cite only. · Sign-up: None

    IA ideas that use it: How fast is a sprinter at each moment? Differentiating 100 m splits · Is there a limit to how fast humans can run 100 m? · Why don't long jumpers take off at 45°?.

24. Olympedia Sport and games
    Olympic results for every event since 1896, with many athletes' heights and masses.

    Format: Web tables · Licence: No reuse licence found: link and cite only. · Sign-up: None

    IA ideas that use it: How fast is a sprinter at each moment? Differentiating 100 m splits · Is BMI's “height squared” right? Evidence from Olympians · Is there a limit to how fast humans can run 100 m? · Why don't long jumpers take off at 45°? · Will women's Olympic times ever match men's?.

25. Project Gutenberg Text and codes
    70,000+ public-domain books as plain text, for letter and word counts.

    Format: Plain text · Licence: Public domain in the US; download a few books by hand (the site blocks bulk robots). · Sign-up: None

    IA ideas that use it: Breaking a cipher with letter frequencies and chi-squared · Can letter transitions tell languages apart? Markov chains for text · Can statistics tell two authors apart? · Finding the key length of a Vigenère cipher with the index of coincidence.

26. Census at School New Zealand Your own survey data
    Anonymous answers from school students to a standard questionnaire, with a random sampler, to compare with your own survey.

    Format: CSV · Licence: Check the terms on the site. · Sign-up: None to browse

    IA ideas that use it: Can you average a 1–5 rating? Analysing Likert-scale survey answers · Do we know how long we spend on our phones? Estimates versus screen-time data · Does the wording change the answer? A split-sample survey experiment · Getting honest answers: the randomised response technique · How big should a school poll be? Sampling from a known population · Walk, cycle, bus or car? Travel to school and distance and 1 more.

27. Census at School USA (American Statistical Association) Your own survey data
    The US version of the same questionnaire, with a random sampler of past responses.

    Format: CSV · Licence: Check the terms on the site. · Sign-up: Check the site

    IA ideas that use it: Can you average a 1–5 rating? Analysing Likert-scale survey answers · Do we know how long we spend on our phones? Estimates versus screen-time data · Does the wording change the answer? A split-sample survey experiment · Getting honest answers: the randomised response technique · How big should a school poll be? Sampling from a known population · Walk, cycle, bus or car? Travel to school and distance and 1 more.

28. UCI Machine Learning Repository Places to find more
    600+ documented datasets (student performance, bike sharing, wine quality and more).

    Format: CSV · Licence: Newer datasets are CC BY 4.0; older ones vary, so check each dataset's page. · Sign-up: None

    IA ideas that use it: How big should a school poll be? Sampling from a known population.

29. Epoch AI data on AI models Places to find more
    Training compute, parameters and publication dates for notable AI models.

    Format: CSV · Licence: CC BY 4.0. · Sign-up: None

30. NASA planetary fact sheet Places to find more
    Orbital period, distance from the Sun, mass and radius of each planet.

    Format: Web table · Licence: US government data; cite NASA. · Sign-up: None

31. WHO Global Health Observatory Population and health
    Over 1,000 health indicators by country: immunisation, mortality, risk factors; plus FluNet and the COVID-19 dashboard.

    Format: CSV, API · Licence: CC BY-NC-SA 3.0 IGO: non-commercial use only, credit WHO, share alike. · Sign-up: None

    IA ideas that use it: How fast did COVID-19 spread at the start? Comparing doubling times · How fast would flu spread through my school? An SIR model · How much vaccination is enough to stop measles?.

32. Johns Hopkins CSSE COVID-19 data (archived) Population and health
    Daily COVID-19 cases and deaths by country, January 2020 to March 2023; no longer updated.

    Format: CSV · Licence: CC BY 4.0 (see the repository's README). · Sign-up: None

    IA ideas that use it: How fast did COVID-19 spread at the start? Comparing doubling times.

33. NCD Risk Factor Collaboration (NCD-RisC) Population and health
    Country estimates of mean BMI, height, blood pressure and diabetes, by sex and year.

    Format: CSV · Licence: Check the terms on the site. · Sign-up: None

34. data.gov.uk Places to find more
    Search across UK government datasets: transport, environment, education, health.

    Format: CSV and others · Licence: Mostly Open Government Licence v3.0; each dataset states its own. · Sign-up: None

35. Kaggle datasets Places to find more
    Community uploads on almost any topic: use only datasets marked CC0 or CC BY that name their original source.

    Format: CSV · Licence: Varies by dataset. · Sign-up: A free account is needed to download

36. UK Data Service Places to find more
    The UK's largest collection of social and economic survey data; a few collections are open, most need a university login.

    Format: Various · Licence: Varies by collection; use only the open ones. · Sign-up: Needed for most collections

Optional browsing tools

Useful for exploring before you download, but never the source you cite.

• Gapminder Tools Population and health
  Animated bubble charts for exploring country data before you download it.

  Format: Interactive · Licence: Explore only; cite the underlying data. · Sign-up: None

  IA ideas that use it: Does money buy a longer life? GDP and life expectancy.

• RareBirthday Population and health
  Browse how common each birthday date is.

  Format: Interactive · Licence: Explore only; cite the NCHS data, not the tool. · Sign-up: None

  IA ideas that use it: The birthday problem, extended: what about birthdays within a day?.

• Open-Meteo historical weather Climate and environment
  Daily and hourly past weather for any location.

  Format: CSV, API · Licence: Data CC BY 4.0; the free service is for non-commercial use only. · Sign-up: None

  IA ideas that use it: How much does a cold month cost my family? · How much does rain reduce bike-share use?.

Collaboration and research platforms

Not data sources as such, but places to find real projects, data stories and ways to question data.

• OSF (Open Science Framework) Collaboration and research platforms
  Where researchers share data and materials, including the large replication projects.

  Format: Projects · Licence: Varies by project. · Sign-up: None to browse

• Zooniverse Collaboration and research platforms
  Real research projects where volunteers classify images and sounds; a few publish their results as data.

  Format: Projects · Licence: Varies by project. · Sign-up: Optional

• NRICH data-handling tasks Collaboration and research platforms
  Tasks that teach you to question and present data well (University of Cambridge).

  Format: Tasks · Licence: Free to use for learning. · Sign-up: None

• Royal Statistical Society Collaboration and research platforms
  The UK's statistics society: careers, statistical literacy resources and competitions for students.

  Format: Articles · Licence: Read and cite. · Sign-up: None

• Data Is Plural Collaboration and research platforms
  A free weekly newsletter of interesting public datasets, with a searchable archive.

  Format: Newsletter · Licence: Each dataset has its own. · Sign-up: None for the archive

Next: an idea that uses real data

78 of our 239 IA ideas have a “Real data for this idea” box naming the primary source, the Our World in Data chart where one fits, and the licence limits. When you have your data, fit a model step by step with IA modelling or test it with IA statistics.