Very cool, thanks Thomas
On Sun, May 8, 2016 at 6:02 PM, Pine W wiki.pine@gmail.com wrote:
Forwarding because this may be of interest to Analytics subscribers as well.
Pine ---------- Forwarded message ---------- From: "Thomas Steiner" tomac@google.com Date: May 2, 2016 01:18 Subject: [Wiki-research-l] [ANN] Wikipedia Tools for Google Spreadsheets To: "Thomas Steiner" tomac@google.com Cc: "public-lod@w3.org" public-lod@w3.org, "Semantic Web" < semantic-web@w3.org>, "Discussion list for the Wikidata project." < wikidata@lists.wikimedia.org>, "Research into Wikimedia content and communities" wiki-research-l@lists.wikimedia.org
Esteemed Wikipedia, Wikidata, Linked Data, and Semantic Web communities[*],
=== tl;dr: Released a Google Spreadsheets add-on called Wikipedia Tools [1] that makes working with data from Wikipedia and Wikidata a breeze. ===
I am happy to release a Google Spreadsheets add-on called Wikipedia Tools [1]. This add-on allows you to work with data from Wikipedia and Wikidata from within a spreadsheet context using custom formulas. Let me motivate the tools with a short example:
You may have heard of Volkswagen's #DieselGate scandal. Is this still a problem for Volkswagen—and if so, where? Google Trends to the rescue? Maybe [2]. But what about global impact? How do people in Korea, an important Volkswagen export market [citation needed😉], refer to the scandal? Turns out they call it 폭스바겐 배기가스 조작 (among probably other options).
With a custom function from Wikipedia Tools, we can safely "translate" from one English (a language that, for the sake of this example, we assume we dominate well enough) Wikipedia article to many other languages (that we do not necessarily dominate):
=WIKITRANSLATE("en:Volkswagen_emissions_scandal") bg Афера на Фолксваген cs Dieselgate de VW-Abgasskandal […] zh 福斯集團汽車舞弊事件
Then, using Wikipedia page views as one (among others) reasonable popularity indicator, for each of these language results, for example for Korean, we can get =WIKIPAGEVIEWS("ko:폭스바겐 배기가스 조작") for the last n days, and plot the results [3] (in practice, you would probably still normalize by size and/or total views of the particular Wikipedia[**]).
There are a lot more custom functions implemented than I could cover in this short example. I have put together a slide deck [4] and paper [5] that go into more detail if you are interested, a demo with all functions is available at [6]. The add-on also has a built-in manual (in Google Sheets, click Add-ons→Wikipedia Tools→Show documentation) and its underlying code is open-source [7].
Please let me know in case of any open question, feature request, or bug. Thanks!
Cheers, Tom
-- [1] http://bit.ly/wikipedia-tools-add-on [2] http://www.google.com/trends/explore?hl=en-US&q=volkswagen+emissions+sca... [3] https://docs.google.com/spreadsheets/d/1PyFq59iEeLWpPQrWDUyU8mlmQrb4GDv2QElm... [4] bit.ly/wikipedia-tools-slides [5] bit.ly/wikipedia-tools-paper (PDF) [6] https://docs.google.com/spreadsheets/d/1sVduZul787O-bRzuy0UKpRl7bkouxwaIOsxX... [7] https://github.com/tomayac/wikipedia-tools-for-google-spreadsheets/ [*] Cross-posted on purpose (http://ruben.verborgh.org/blog/2014/01/31/apologies-for-cross-posting/), please choose your reply options accordingly. [**] This is a simple example for illustrative purposes, I do _not_ claim it is an accurate popularity prediction, nor do I mean to bash Volkswagen.
-- Dr. Thomas Steiner, Employee (http://blog.tomayac.com, https://twitter.com/tomayac)
Google Germany GmbH, ABC-Str. 19, 20354 Hamburg, Germany Managing Directors: Matthew Scott Sucherman, Paul Terence Manicle Registration office and registration number: Hamburg, HRB 86891
-----BEGIN PGP SIGNATURE----- Version: GnuPG v2.0.29 (GNU/Linux)
iFy0uwAntT0bE3xtRa5AfeCheCkthAtTh3reSabiGbl0ck0fjumBl3DCharaCTersAttH3b0ttom hTtPs://xKcd.cOm/1181/ -----END PGP SIGNATURE-----
Wiki-research-l mailing list Wiki-research-l@lists.wikimedia.org https://lists.wikimedia.org/mailman/listinfo/wiki-research-l
Analytics mailing list Analytics@lists.wikimedia.org https://lists.wikimedia.org/mailman/listinfo/analytics