Un projet de

com a colaboração da

avec le soutien de

com o apoio de

no ambito do

Base de données réalisée avec le soutien de

banco de dados feito com o apoio da

INDICATEURS DE LA PRESENCE DES LANGUES DANS l’INTERNET : accès en français

INDICADORES DA PRESENÇA DAS LÍNGUAS NA INTERNET : acesso em português

INDICATOR OF THE PRESENCE OF LANGUAGES IN THE INTERNET : access in English

OBDILCI.ORG | | CONTACT

View Cybergeography evolution

Credits

The project of OBDILCI about creating indicators of the presence of languages in the Internet started in 2017, have reached version 3.2 in April 2023, and version 4 in May 2023.

The project has been conducted by OBDILCI with the collaboration, starting from version 2, in April 2021, of the UNESCO Chair on Language Policies for Multilingualism. The project has been funded by Organisation de la Francophonie, for previous studies, versions 1 and 3 and by the Brazilian Ministry of Foreign Affairs via Instituto da Lingua Portuguesa for version 2. The realization of the database access and the publication of its methodology in Frontiers Research Metrics and Analytics, has been funded by Délégation générale à la langue française et aux langues de France, from France Ministry of Culture, together with the Permanent Delegation of Brazil to UNESCO.

Sources

The method behind the unprecedented production of indicators of the presence of languages in the Internet

For demo-linguistic data, Ethnologue, a proprietary source, updated once a year.
For percentages of persons connected to the Internet per country, ITU and World Bank, both public sources updated once a year.
For figures about languages or countries on the Internet, a large variety of public direct sources (such as Imminent T-Index) or some indirect sources (such as similarweb.com applied to a series of websites) which are all listed in https://www.frontiersin.org/articles/10.3389/frma.2023.1149347/full#supplementary-material.

Indicators

The indicators produced by OBDILCI are accessible under CC-BY-SA 4.0 license, in Excel files from https://obdilci.org/Results or in the form of a database query in https://obdilci.org/Base. The results from Version 3.0 are fully described in the peer-reviewed, open data article Resource: Indicators on the Presence of Languages in Internet. The indicator's figures are displayed with 2, 3 or 4 digits after the coma, depending on the context; however, it must be understood that the contents indicator computed by the model is obtained within a rather large confidence interval (±20%) and therefore differences lower than 20% are not significant for comparison between language's contents.

WHY?

The queries are expressed via ISO CODE 639-3 not via the language name.
The reason is that many languages have different names and/or different orthographs and additionally some of those names are common with other languages. The non-ambiguous way to select a specific language is thru ISO 639-3.

HELP

If you do not know the code of the language you want to query try first introducing the first 3 letters of the English name of the language, it may work for several languages. If not, search for the language name in https://en.wikipedia.org/wiki/Wikipedia:WikiProject_Languages/List_of_ISO_639-3_language_codes_(2019). As a last resort, you will find the code using your favorite search engine with a request such as: “language xxxxxx iso 639-3”.

Local Name(s)

The English names of languages corresponding to ISO-639 are those used by Ethnologue. For translation of the names to French and Portuguese the process starts with https://en.wikipedia.org/wiki/Wikipedia:WikiProject_Languages/List_of_ISO_639-3_language_codes_(2019), using the English Wikipedia article corresponding to each associated language and the translation of those articles using Wikipedia's other languages link to the French article and from there to the Portuguese article.

In case no French article is indicated, the same process is used from the Portuguese article. In case no corresponding translation is found further research is realized. In case no translation is secured the English name is kept for translation.

For the localized name the first source is the autonyms field of https://www.ethnologue.com/language/isocode and the second source is the Wikipedia article of the corresponding language.

For macro languages, except a few exceptions of macro-languages grouping a very large number of languages, all forms of the corresponding languages are searched for and listed separated by a "+".

Very few localized names are not found and are left empty (this may be the consequence of some Unicode conflict or lack of standard).

If somebody wants to help fill this gap, please use the contact form and explain your source.

L1+L2 SPEAKERS

Speakers of this language as first or second language. (source: Ethnologue). Note that the percentage is not computed over world population but over the total of L1+L2 speakers (X).

L1+L2 CONNECTED SPEAKERS

Speakers of this language as first or second language who are connected to the Internet out of the total connected speakers. (Source: elaboration made by the Observatory from Ethnologue and ITU + World Bank sources).

PERCENTAGE OF L1+L2 ENGLISH SPEAKERS CONNECTED TO THE INTERNET

Percentage of speakers of this language as first or second language connected to the Internet. (Source: elaboration made by the Observatory from Ethnologue and ITU + World Bank sources).

CONTENTS

This is the percentage of Web contents in this language (note that as Humans, and probably still more, websites may be multilingual (source data computed by the Observatory model from multiple sources; see the Obdilci website for detailed explanation of the method).

VIRTUAL PRESENCE

The virtual presence is the ratio of D over A. A value of 1 means a fair virtual presence. A value over 1 means a good virtual presence, over the real-life presence; A value below 1 means a bad virtual presence below the real-life presence. The experience has shown that this value hardly gets out of the window 2 - 0.5.

CONTENT PRODUCTIVITY

The content productivity indicator is the ratio of D over B, it measures the propension of connected speakers of this language to produce contents. A value of 1 means a fair productivity. A value over 1 means a good productivity; a value below 1 means a bad productivity.

Multilingualism factor

This means that more or less of world population speaks more than one language (note that a person is counted as L2 as many times as that person speaks different languages).

TREND

Trend measures, as a positive or negative %, the change of contents % since the last measurement. Note that if the value is left empty for some languages it is because they were only included in the latest version, so there is no previous value for comparison. Note also that extreme changes are most probably due to data correction on statistics, either demo-linguistic or country connectivity to the Internet (not in a real sudden evolution of contents for those languages).

L1 speakers

The first language (often noted L1) is the mother tongue. Although it takes some months for babies to start speaking they are computed with their mother tongue and the world total of L1 speakers is then equal to the world population. L2 are all the second languages spoken by a person, supposedly at a level of fluency sufficient to maintain conversations on a wide variety of subjects. Not all persons have second languages, but obviously the world total of L1+L2 speakers is much higher than the world population and this figure must be used to compute percentages concerning L1+L2 populations. Many studies show that persons usually prefer to use their L1 language on the Internet but can also make use of their L2 language(s) in a variety of situations. This is why all the indicators are produced on a L1+L2 basis. Note that websites, as humans (and probably more than humans), can be multilingual; not paying due attention to that fact induces important biases to linguistic statistics about the Web.