acWaC-EU is a 90-million word corpus of web pages in English crawled from the websites of European universities. The universities sampled are based in countries where English is or is not a native/official language, hence this (monolingual comparable) corpus affords comparisons of native and lingua franca (ELF) varieties of English in the institutional academic domain.

Each text in the corpus (corresponding to a single web page) is annotated with rich metadata, including:

• URL from which each text was downloaded and “level” in the site structure at which it was found (from 0 to 2, where 0 indicates the homepage);
• variety of English (native/ELF);
• name of the university which published the page and its rank according to Webometrics world rank of universities;
• country where the university is based, European Union membership (yes/no), and language family of the official language spoken (e.g. France/Romance, Norway/Germanic, Russia/Slavic, etc.).

This page provides information on acWaC-EU and the scripts that were used for its construction.

In the near future the corpus will be made available as a set of n-grams, along the lines, e.g. of the Google Books N-gram dataset or the Rovereto Twitter N-Gram Corpus. Stay tuned!

When referring to the acWaC-EU corpus please cite:

Ferraresi, A. and Bernardini, S. (forthcoming). The academic Web-as-Corpus. In Evert, S., Stemle, E., and Rayson, P. (eds). Proceedings of the 8th Web as Corpus workshop (WAC-8), Lancaster, UK.

The table below presents detailed information on the contents of the acWaC-EU corpus.

The plot below shows the number of tokens (in %) for the “main” language families featured in acWaC-EU (i.e. those with 500K+ tokens).

Here you can download the main scripts that were used to build acWaC-EU — more detailed documentation will follow (hopefully soon):

• the script to obtain the list of all European universities from Webometrics (the script requires countries_langs.info, included in the archive; the archive also includes a new script to obtain the list of all universities worldwide);
• the script to look for English-language homepages;
• the postprocessing scripts (adapted from the scripts developed within the WaCky project).