2 =======================
4 __ https://en.wikipedia.org/wiki/Main_Page
6 .. figure:: seattle_redirect.png
10 Example of a redirect displayed in the `Seattle, Washington page on
11 English Wikipedia`__ which is a redirect to `Seattle`__. Note the
12 small redirect notice below the title.
14 __ https://en.wikipedia.org/wiki/Seattle,_Washington
15 __ https://en.wikipedia.org/wiki/Seattle
17 In wikis, **redirects** are special pages in that silently take
18 readers from the page they are visiting to another page in the
19 wiki. In the `English Wikipedia`__, redirects make up more than half
22 Different Wikipedia data sources handle redirects differently. For
23 example, `the MediaWiki API`__ will automatically "follow" redirects
24 but the `XML database dumps`__ treat redirects like normal
25 articles. In both cases, redirects are often invisible to researchers.
27 __ https://www.mediawiki.org/wiki/API:Main_page
28 __ https://meta.wikimedia.org/wiki/Data_dumps
30 Because redirects constitute a majority of all pages and see a large
31 portion of all traffic, Wikipedia researchers need to take redirects
32 into account or their findings may be incomplete or incorrect. For
33 example, the histogram on this page shows the distribution of edits
34 across pages in Wikipedia for every page, and for non-redirects only.
35 Because redirects are almost never edited, the distributions are
36 very different. Similarly, because redirects are viewed but almost
37 never edited, any study of views over articles should also take
38 redirects into account.
40 .. figure:: edits_over_pages.png
44 Histograms of pages in the article namespace based on the number of
45 edits on a log scale. The red histogram includes all pages in the
46 article namespace (N = 9,729,989). The blue graphs includes only
47 pages that were not redirects at the point of data collection (N =
50 Because redirects can change over time, the snapshots of redirects
51 stored by Wikimedia and published by Wikimedia Foundation are
52 incomplete. Taking redirects into account fully involves looking at the
53 content of every single revision of every article to determine both
54 when and where pages redirect.
56 Much more detail can be found in `Consider the Redirect: A Missing
57 Dimension of Wikipedia Research`__ — a short paper that we have written
58 to accompany this dataset and these tools. If you use this software or
59 these data, we would appreciate if you cite the paper:
61 *Hill, Benjamin Mako and Aaron Shaw. "Consider the Redirect: A Missing
62 Dimension of Wikipedia Research." In Proceedings of the 10th
63 International Symposium on Open Collaboration (OpenSym 2014). ACM
66 __ http://mako.cc/academic/hill_shaw-consider_the_redirect.pdf
68 Generating Redirect Spells
69 =============================
71 Generating redirect spells from an MediaWiki XML dump involves two steps:
73 1. Searching the full text of every revision of every page in a dump to
74 determine if any given revision is a redirect.
76 2. Using the results of (1) to generate a list of "spells" that describe
77 periods of time that articles in a wiki redirect to other articles.
79 We have `publicly released software in Python and R to do these two
80 steps`__ under the `GNU GPL version 3`__. The software is designed for
81 people already comfortable with working with MediaWiki XML dumps and the
82 tools and software necessary to do this.
84 __ http://projects.mako.cc/source/?p=redirect-tools
85 __ http://www.gnu.org/licenses/gpl-3.0.html
87 You can download the software from our git repository like::
89 git clone git://projects.mako.cc/redirect-tools
91 Detailed documentation on how to use the software is in available in `our
97 =========================
99 In `our paper`__, we present an analysis of redirect data from English
100 Wikipedia in the dump created in October
101 2012. You can download `the dump files we used`__ from `the Wikimedia
102 Foundation dataset archive`__. Because generating these dumps can be
103 computationally intense, we have published the output of the software
104 above run on the this dump. This includes 9,277,563 redirect spells that
105 our software identified and is the dataset used in the paper.
107 You can download the dataset in the following formats:
109 - `RData (240MB)`__ — Suitable for use in `GNU R`__
110 - `bzip2 compressed tab separated values (178MB)`__ — Suitable for use
111 in other languages and statistical packages.
113 __ http://mako.cc/academic/hill_shaw-consider_the_redirect.pdf
114 __ http://dumps.wikimedia.org/enwiki/20121001/
115 __ http://dumps.wikimedia.org/
116 __ enwiki_201210-redirect_spells-v1.RData
117 __ http://www.r-project.org/
118 __ enwiki_201210-redirect_spells-v1.tsv.bz2
123 For details about the dataset, why it is important, and for examples on
124 how it can be used to come to better findings in Wikipedia research,
125 please read `the companion paper`__.
127 __ http://mako.cc/academic/hill_shaw-consider_the_redirect.pdf
129 If you notice issues or bugs in our data or `code`__, contact `Benjamin
130 Mako Hill`__ or `Aaron Shaw`__.
132 __ http://projects.mako.cc/source/?p=redirect-tools
133 __ http://mako.cc/contact/
134 __ http://aaronshaw.org/
136 Patches and improvements are welcome! Details on `how to produce and send
137 a patch using git are online`__.
139 __ http://projects.mako.cc/source/
144 ⓒ Copyright `Benjamin Mako Hill`__ and `Aaron Shaw`__ :: `Creative Commons BY-SA`__ :: Updated: Thu Jul 3 13:22:29 PDT 2014
146 __ http://mako.cc/academic/
147 __ http://aaronshaw.org/
148 __ http://creativecommons.org/licenses/by-sa/4.0/