speakrow.

About Speakrow

Speakrow is an independent site about the languages and geography of the world. It is built on one dataset of 203 countries, 411 languages and 874 country-language pairs, each carrying a share of the population and an official-status flag. That join between countries and languages is what the whole site is for: it is common to find a list of a country's official languages, and much rarer to find what people there actually speak.

What is on the site

The linguistic reach tool is the original idea: pick the languages you speak and see what share of the world's population you could hold a conversation with, drawn country by country on a map. The language reach tables answer the same question in reverse, ranking languages by how far each one gets you and how much a second or third one adds.

The country pages give the same data one country at a time. The daily games and visualizations are ways into the same dataset from different directions.

Where the data comes from

Language shares and official status come from Unicode CLDR, which estimates, for each country, the share of the population speaking each language. CLDR counts both first-language and additional speakers, which is why English reaches far more people here than its native-speaker count would suggest.

Country facts, including capitals, area, land borders, currencies and alternate names, come from the mledoze/countries dataset under ODbL 1.0. Development indicators such as GDP per capita and life expectancy come from World Bank Open Data under CC BY 4.0, and each value carries its own reference year because countries report in different years. Map shapes are Natural Earth. Language snippets in the guess-the-language game are from the Universal Declaration of Human Rights. Flag images are served by flagcdn.com.

How the reach figure is calculated

For every country, the population shares of your selected languages are added together and capped at 100%, then multiplied by that country's population. Summing across all 203 countries gives the number of people reached.

The cap is the part that matters. In a country where 90% speak Hindi and 40% speak English, those two languages do not reach 130% of the population. They reach, at most, everyone. This is why a second language adds far less than its raw speaker count implies, and why the biggest gains come from languages spoken where your existing ones are not.

What these numbers are not

They are estimates, not census counts, and they should be read as approximations. Language demographics are genuinely hard to measure: definitions of what counts as speaking a language vary, many countries do not ask, and several of these figures are modeled rather than observed.

"Reach" means a shared language exists, not that conversation is guaranteed. Someone counted as an English speaker in a country where English is a second language may have very different fluency from a native speaker. Where a value is genuinely missing from a source, it is shown as no data rather than estimated or interpolated.

Who makes it

Speakrow is an independent project, built and maintained by one developer rather than a company or an institution. It has no funding, no editorial board and no affiliation with any of the data providers named above.

The developer's personal site is marcosburlino.com. To report a data error or a bug, use the contact page, or the menu button in the top right of any page.

Common questions

Is Speakrow free?
Yes. Every tool, game and visualization is free, with no account and no paid tier. The site carries a small number of ads to cover its costs, placed below content rather than over it.
Can I cite Speakrow?
Yes, and please include the reference year where a page states one, because World Bank indicators are published for different years for different countries. The underlying datasets carry their own licences, named on each page, which apply to the data itself.
I found a mistake in the data. How do I report it?
Use the contact page, or the menu button in the top right of any page and choose Report a Bug. Both go to the same place. Corrections to country and language data are welcome and genuinely useful, since much of this data is estimated rather than counted.